Beware Bait and Switching WP Engine! A Case Study in Devolution

One of my side projects is keeping tabs on devolution, which I first discussed in January in Devolution: Welcome to the World Where Things Don’t Work Well.

I’m going to go a bit Consumerist and discuss a recent frustrating case. Some readers might have better surmises than I do as to how this fits into the overall pattern of people not building organizations and products like they used to.

The example is my experience with WP Engine. By way of background, WP Engine is a specialist WordPress hosting service. WordPress is sufficiently fussy that it is prone to breakage and performance issues. As a result, there is a clutch of WP-oriented webhosts who cater to WP users (and please don’t tell me the issue is the hosting, not the software. The very existence of WP-focused hosts who charge a hefty premium to standard big iron hosting should be sufficient proof that this is a real issue). I will spare readers the discussion as to why this is probably what I need, but trust me, this is not something I’ve come to based on a whim.

WP Engine was highly recommended by a couple of readers. I’m not sure quite when I called them (November or December-ish) but the first sales call was unbelievable. I had looked at their service tiers (they have three tiers where they show pricing, and the next is “Premium – call for a price” and thought I fit into the priciest tier below Premium but wanted to be sure (as in I did not want to migrate under an incorrect assumption about what I was getting into). I went through basic parameters on my site (traffic, # of posts in the database, frequency of posting and commenting, since those all hit the database, which is the most problematic part of WP). The rep’s attitude was she wanted to get off the call with me as quickly as possible. She was not interested in my questions re what tier I fell in or the nature of their support. She kept referring me to the site when I’d already looked at it.

I sputtered at Lambert about the call. The message was they didn’t want my business. I should have taken that as a warning.

I then had a misadventure with another WP specialist hosting service, a relatively young one. I had scheduled a migration. Turns out despite agreeing a price, signing up, and setting a date, the migration was somehow NOT scheduled on their end. And it appeared this was the result of too many things converging on the CEO and him being chronically overloaded. Now I know that space. I live it. But that is death in a service business. And indeed, someone who did migrate to them a little after my aborted migration had a disaster and exited within weeks.

So I decided to try WP Engine again, since there aren’t a lot of options (WP VIP at the time was another choice, but I hated the idea of enriching WP for the crappiness of their software by giving them hosting business). This is January. I have a fairly decent conversation with one salesman who says I’m in their highest tier below premium and says they offer 24/7 software support which I had figured they did but I wanted confirmation. He also said I could self migrate or use a service called WP Valet for a fixed price of $199. I decided to talk to WP Valet.

WP Valet responds to the e-mail quickly, and said the base price for a migration is $199, which is not exactly what WP Engine had said. So about a week of discussion with WP Valet ensues where they check my plugins to make sure they are all OK with WP Engine and have me give them info like DB size, raw dump. We discussed doing some housekeeping to reduce the DB size. I had to upgrade in order to migrate which they said they would do for an additional $50.

So far this is taking longer than I’d expected but nothing is horribly amiss. But given the extra charge, I figured I’d better talk to WP Engine again to make sure I haven’t missed something. I again have a rep say based on the stuff I told them, starting with 1.4-5 million pageviews a month, that I was in the third of their four pricing tiers. And mind you, in none of these three calls did they ask questions that I did not answer.

But I did learn that the software support was not what I thought it was. It turns out 24/7 software support didn’t mean they’d fix problems if the site fell over, which is what I had discussed with the first two reps and they had indicated was what they had. And that was the feature I wanted most of all.

The latest rep said they would help me troubleshoot if the site fell over. Visions of dealing with Verizon when my service goes down flooded my brain. This seemed to be a big difference from what I had understood previously, but at this point I was psychologically committed to going ahead.

Some more back and forth on mechanics with WP Valet follows. I’m now puzzled that nothing is set up on the WP Engine end regarding billing, which seems pretty bizarre with a migration due to start. So I call WP Engine again. I go through the same song and dance as before with yet another salesperson. This salesman tells me I am in the Premium tier. The minimum price is double what I’d been told up to then.

Needless to say, I was outraged and canceled the migration.

Two and a half months later, I’m a beached whale on a web redesign because my current support guy can’t set up a workable dev space for my designer. This has been going on for months. This isn’t uncommon for WordPress. It apparently needlessly complicates the creation of dev sites or staging sites, and the tradeoff of letting site admins install plugins or tweak code is that you end up fighting with the server over file permissions.

But this seemed to bring the need to get into a setting with people who knew how to deal with it. And my designer had recently migrated a very fussy site (extremely busy ads which they serve, plus 31 plugins v. my mere 11, at 500,000 page views a month) and she said once you got past the salesmen, the tech guys were really good.

So I called a fifth time.

I was told their minimum for the Premium tier was $100 a month higher than I was quoted a mere 11 weeks earlier. The salesman said this was impossible (ie, I’m a liar) since they’d increased prices at the beginning of the year. I said no, my designer had migrated a site over this year and was paying the same price I’d been quoted. I gave particulars but he wasn’t listening.

I sputter to my designer. She tries to working this from the tech end, since the site she migrated over is much higher maintenance than mine. Her messages got more pointed:

To now say she can’t have that price because she could not get through the sluggish sales process in time is something that sounds Kafka-esque in its rigid technical interpretation mired in procedure and reflects poorly on how the company does business. This has been the problem from the beginning. This is what you and I discussed on the phone, and because they are also very low maintenance you told me this would not be a problem. It is still a problem. Susan has been jerked around since moment one in January when I first brought her to WP Engine. I am simply asking that after all of the run-around to get to this point, she be given the price she was told.

I can’t begin to fathom how anyone can conduct business like this. I was baited and switched twice on pricing. And then I get all sorts of caveats while we are seeing if they will relent that were totally absent in the previous conversations: “Any quote you were given couldn’t have been more than a ballpark…” Oh, and this even better: “At WP Engine, we want to be able to stand behind the promises we make to you — and a quote is pretty much a promise.” So that’s operative now? Why not in my previous four conversations when I was pressing for answers?

Now this may not seem like a big deal, but I’ve spared you a lot of detail. Twice I geared up for a migration (the tech guy was optimistic he’d be able to persuade the sales guys), which means rearranging schedules (you need to check that everything came over, and I had lined up both Richard Smith and Lambert to help, plus got my current site guy to take some measures to trim the DB, and I did some housekeeping too). So I wasted a lot of cycles for a company that can’t even do the most basic thing in sales, which is to give realistic pricing to an interested customer that will pay. And more important, they demonstrated repeatedly that they are not people of their word.

So I don’t care how good their back end is. I can’t trust people who do business like that. And the big lesson for me is I should have listened to my initial instincts.

But from the bigger perspective, what can a company be thinking that has a sales operation like this? I can only fathom that they believe they have a captive audience and can be pleasant superficially but substantively jerks in the sales process. And that’s not entirely inaccurate. When I came back the third time. WP VIP, their big competitor, had dropped its tier for folks like me and will only take on monster clients.

But even people who don’t currently have a lot of options will still go somewhere else rather than deal with this sort of treatment. Remember, WalMart, who was confident that its low prices would always prevail, is managing to drive customers to higher-priced Costco and Target. In my case, I’m back in a holding pattern and a tech person surfaced who can sort out the immediate problem, the permission woes. So you might eventually see a spruced-up NC after all.

Print Friendly, PDF & Email


  1. Clive

    From my experience (big software implementations in the financial services industry) the entire IT sector is a rotten borough. All the players, consultants, sales force, technicians/developers, system integrators, testers, project management — you name it — are trapped in the same system. Regardless of their ethics or talent, the rules are the same for all actors.

    What are then the rules ? Well, as had been commented on previously, one of the worst sorts of “markets” (which is where the whole “free market” flim-flam comes completely unstuck)is where you have the (in)famous “market for lemons”. Put simply, you cannot tell in the IT industry who is good, who is bad and who is just plain ugly. Price is, alas, no guide either.

    The whole industry (that’s much too nice a word for it all) reminds me of patent medicines before the FDA, accountancy before GAAP, the stock market before the Kennedy commission in the 1930’s (cough, what happened to that good work?), bread making before the Weights and Measures Act (that’s an English example, am sure there’s similar in the US) etc. etc. etc. In other words, it’s unregulated, people can purport to sell anything they like regardless of their ability to actually deliver on the sales pitch. All the consumer has to fall back on (and this applies from the single customer like Yves’ predicament) way up to the big corporate account is basic consumer protection and contract law. Which means, most of the time, people aren’t prepared to have the hassle and uncertainty of litigation if things go awfully wrong.

    A further nuance is, and this ties in with de-evolution, is that when you scratch beneath the surface of many (if not all) suppliers in this industry, there’s a huge scarcity of trained, competent labor. IT is complex. It takes years and years to learn what is actually a craft trade. What should happen is that enterprises should be training up college grads and turning their technical theoretical knowledge into real-world experience and knowledge. But guess what Virginia, no-one wants to train up new talent (we’ve done that subject before too here on NC) for a variety of reasons (e.g. the might if you’re horrible to them in the workplace want to, shock-horror, leave so you’d be better off in the long run being a bit nicer to them… try selling that one to the c-suite).

    So, in summary (sorry, long comment, am almost done) this is a key industry, it can’t/won’t enforce its own minimum standards and so should have some regulation and/or a professional guild which you can get chucked out of if you misbehave.

    Chances of that happening ? About 0%. So, lucky consumers, you get to gamble with either your own or your stockholders money on the crap table layout that it the IT business.

    1. Kaz Augustin

      Absolutely Clive!

      Both my husband and I are veterans in the IT industry, but having people with degrees in Music or History somehow leapfrogging over us to manage *technical* projects is a trend that makes us want to vomit. IT has become a circus of appearance over substance and there’s not a goddamn thing any of us old-timers can do to turn the clock back. After all, an aspiring IT technical manager with a degree in flamin’ Biology is not going to ask for as much money as an IT technical manager with several degrees in Computer Science, is s/he?

      Sympathies and welcome to the club, Yves. Take a number and join us at the bar.

      1. DanP66

        I agree whole heartedly.

        Part of the problem is the constant drive to push down costs.

        1. We let go experienced people because they are expensive and replace them with “good enough” people that are cheap. Look at what CGI & CSC have been doing to their staffs lately as an example.

        2. Everybody builds for efficiency ( of a kind ) and nobody considers the robustness of what they build be it a process, a network or an application. I bet part of Yves problem is a result of bad processes, bad communications and bad training as much as anything else.

        3. Nobody cares about the downstream effects of items one and two all they care about is dollars TODAY.

        1. Banger

          As a veteran of many IT projects the more money you save by hiring budget staff the more money you spend–it’s an iron law for any slightly complex project. I think that is even more the case today when there exponentially more directions to go in.

    2. McMike

      Indeed. I’ve had many conversations about this with small producers and activists in the local food space. (Along the lines of the Maine farmers’ links post from yesterday). Yes, current fod regulation clearly favors the oligarchy at the expense of small producers and innovation. Yes, the large producers game these regulations to deceive consumers.

      But also… the reason we have regulation and standards is precisely because at one point, the variety of unregulated industry players was ripping off, confusing, sickening, and abusing consumers and generally gumming the mechanisms of commerce by threatenign to precipitate a consumer strike and threatening to Gresham-ize the legitimate players.

    3. ambrit

      Yer, indeed! One of my sons in law is a code writer at the North American branch of a European firm that designs navigation systems for large commercial shipping. He is very good at what he does, and, how felicitous, likes what he does.
      Perhaps this firm is the exception that proves the rule. They treat their employees very well, because, as one of their people told me when I met him; “If you screw up this job, you’ve just sunk a tanker. And the consequences are dire indeed!”
      The above is to highlight the idea that the level of consequences attendant to criminality and incompetence are crucial to the maintenance of quality. We seem to be seeing the results of the absence of consequences everywhere nowadays, don’t we.

      1. McMike

        Tell your son to get some equity now. That company is a prime takeover target.

        Any company that builds and stands behind a quality product, has competent staff that are treated well, and gets a reputation for quality is nothing more than a golden goose just BEGGING to be emptied and then strangled by the masters of the universe (with thier free taxpayer subsidized capital and major tax preferences).

        That’s the sick part: we (the people) are finanicng our own demise – in a system that not only rewards, but fairly demands self-consuming piracy. We build great merchant ships and then send them out into pirate infested waters with a big sign on the back that says; RAID ME.

        No convoys of regulation (or even national ethical principles) that protect our vital industries, no sir, we send them out into wolf packs of hedge fund u-boats on their own, with their rudders locked and a bounty on thier hulls.

  2. Synopticist

    “This seemed to be a big difference from what I had understood previously, but at this point I was psychologically committed to going ahead.”

    That’s”lowballing”. A classic, un-ethical sales techniques which is as old as the hills.

    1. McMike

      re lowballing. Moreso because with tech it is even more sticky, very real time and money is devoted to preparing for and making the transition. Pulling the plug means flushing real sunk dollars.

  3. YankeeFrank

    This sounds like what happens at many organizations today — in short, there is no organization. Turnover is constant, training is sporadic and thin, communication and coordination is almost non-existant, and policies and oversight by the ones who make all the money fails before it even begins.

    Software organizations are notorious for salespeople that will promise you everything without checking to see if its possible, and then dump you on the techs who don’t have the infrastructure to provide what was promised.

    I may be repeating you Yves, but to me its indicative of upper management/ownership that are only interested in profit-taking and disregard everything else, including paying enough to hire talent, treating the talent well enough to keep it, and foregoing their expensive perks and bonuses at least until a sound foundation is developed.

    The sales/tech divide is key here as its usually the sales side that has the relationship with upper management while tech, even in a tech firm, is treated like a backwater where employees are seen as fungible.

    The other problem that I see is the sorry state of software development and integration. Software engineers and architects love their toys too much, and are more interested in having fun than keeping it simple. Add to this the fact that they have little reason to give a crap about the long term stability of the systems they work on (a less culpable version of you’ll be gone I’ll be gone) and a tactical mindset of putting out fires rather than achieving an overarching strategic vision.

    It also sounds like wordpress, as you say, is quite the dog. What I find interesting is that this stuff is not rocket science. But building a business on a shoddy framework is a recipe for disaster. I worked for a hedge fund servicing company for a while a few years ago. They were run by a few guys who had built a very successful company of the same type and had decided to do it again, but this time they would use “cutting edge” tech — “agile” development using python, wrapped around a core set of financial analytics with a thin .net display layer. Python is great, except its not the fastest language, which can be okay because its easy to embed C or C++ to handle computationally intensive parts of the code. However, the analytics library, written in python, was totally out of their control, and the .net view layer in front of the python layers was also a dog performance-wise.

    The business model was to take massive trading data dumps from hedge funds around 3-4pm every afternoon, crunch the numbers asap and get the results back to the client so they could leave at 5. Now, number one concern for such a business is performance. It has to be super fast. Well, with a clunky and slow front end built on top of python analytics the system was starting from a weak position. It could’ve possibly succeeded of course, if done right, but the “agile” development true believers took one of the agile principles, which is to avoid over-engineering a system by building incrementally and ignoring issues like performance in order to keep things simple, at least until a prototype was done. Of course, the design decisions made in the prototype stages tend to lock you into a design later on, unless you do the hard work of redesigning from the ground up, and what’s the point of a prototype if you are going to throw it out and start over? Well, there are reasons but instead they fell in love with their prototype code and ignored performance until the system was built.

    Another fetish was to attempt a RESTful architecture, which basically means that you don’t cache calculation results, because that creates “state”, which can become stale and adds complexity (http and the web are, or were, examples of RESTful services — everything gets loaded from a server, exists as static state in your browser, and then gets thrown out and replaced by other static state when you click a link). But without caching results, calculations may have to be run many times at different points in order to get data that other calcs can rely on.

    The upshot was that the system didn’t work, and only sort of worked if they threw tons of hardware into the mix so that all the calculations could run in parallel, which is expensive but even more importantly, fragile and inflexible. And they just couldn’t get enough hardware into the mix because at some point your network bandwidth becomes an issue (they had so many separate processes piping data into each other it was a nightmare) and your processes become “I/O bound”, meaning they spend way too much time waiting on and reading data from other processes…

    This is a long way of saying that software development is complex, but it becomes way more complex when no attempt is made to be practical, and when engineers spend way too much time and attention conforming to some ideal paradigm than to the concrete tasks at hand.

    Its ironic because agile was an attempt to simplify design. It has many great ideas, but engineers are often true-believers, putting “faith” where it doesn’t belong and leaving common sense in the gutter.

    It sounds like wordpress suffers from either this disease, or the even more common disease whereby a solid core system is destroyed by adding more and more functionality on top that eventually destabilizes the core. This phenomenon is exacerbated of course by high turnover and loss of institutional knowledge.

    Software engineering has to become a lot less sexy, and a lot more like designing bridges, before these kinds of nightmares go away.

    Redesign takes time and money, adds up front risk and takes courage to demand. However, without it, you get my hedge fund servicer fiasco and systems like wordpress, which simply punt the risk down the road, at which point its a much hairier mess and generally impossible to fix.

    Sounds like our financial system actually…

    1. wunsacon

      If that’s how they operate, how good are they on the backend? … Do you backup the site once a week to some other system such as your home computer?

      >> When I came back the third time. WP VIP, their big competitor, had dropped its tier for folks like me and will only take on monster clients.

      I can understand. Running this stuff isn’t trivial. The prices you quote sound cheap, unless everything’s automated. And automating stuff isn’t cheap either, unless it’s one-off and it’s “okay” for it to break at the next upgrade of WP.

    2. Synopticist

      Sorry YF, i didn’t pick up some of those technical details.
      Would you mind repeating it?

    3. fajensen

      Software engineering has to become a lot less sexy, and a lot more like designing bridges, before these kinds of nightmares go away.

      Well, if it is any help, I notice that the software people I know drive increasingly more shitty cars ;-p

      A very common problem, is that, contrary to the building of aircraft and bridges, noboby bothers to hire an architect or a system engineer!

      Layered on top is that software is a transient business, especially in anything web-related: It is assumed that a large enough bunch of 18-25 year-olds fueled by free Pizza, Coke and (the dream of) Stock Options can crank out enough code and re-installs to keep the business running long enough for the CEO to sell the business and cash out.

      Then there is talent, which is not scalable. I worked with software for 10 years before going back to electrical engineering, I am a decent developer, which means that someone better than that, can write the code that I spend maybe 6 weeks writing in 2 weeks, an expert will write the same within 6 days – while also publishing a conference paper and a patch for the Linux Kernel. That expert costs money, which dilutes CEO-value, so most companies will resort to a More Monkeys hiring strategy.

      Contrary to common management theory The Code Monkeys are not entirely stupid – most of them understand the “Transient Business Game” well enough to adopt those tools, frameworks and methodologies that “pad” the CV most effectively. So, if a software business lives more than he intended few years, there will be layer upon layer of dead frameworks in the code, mixed with whatever was hot at this years developer conferences. This is accepted because the old code “works”, “new features” is what bringeth the bacon and the new GUI is very much shinier than the last version – which is what the customers see (apart from the charges on their credit card).

      That is the sorry state of most IT-shops, unfortunately!

      Did i mention that Margins are evaporating? We can buy hosting for a lot less than the electricity for runnning a computer 24/7/365 costs.

      In the not so far future, give it 2-3 years, Amazon EC2 & Co. will have gobbled up all the “own-kit service providers” and left no margin for the emerging “build it on top of the cloud”-shops.

      I think that the best hosting available “these days” is the “no extras” techie places, where all one get is “a computer and a database” – and one pays someone directly to set it up. “Rugged selv-reliance” is where we are all going anyway ;-)

    4. banger

      I don’ think her needs would ever get to the complexity you describe. Having been involved in a project that ran into the sorts of problems you describe (I ran away as fast as I could from that one) I can vouch for the headaches.

      Things change so fast and there are so many alternative methods depending on scale to choose from and the people you have all may have different backgrounds–it’s not like the old days when everybody knew the same languages and used the old structured programming approach or whatever.

      What we need is programmatic approach to programming, i.e., programs that will evaluate modules/libraries and choose the appropriate ones for the task–it’s too confusing for humans.

  4. kemerd

    why not just opt for server space and do all the web serving stuff on your own? 5 million hits a month is not a big deal. In fact, it can be handled (with a decent connection) from your home with an average server. No fancy hardware required

    1. wunsacon

      Agree re performance. But, maintaining a server running WP and connected to the internet is non-trivial, IMO. Security, backups, maintaining software and its dependencies.

    2. Yves Smith Post author

      I have no time. You seem to forget that I am chronically time stressed and need LESS, a LOT less, on my plate. The server stuff should be like the electricity. It should be there, work, and be someone else’s problem.

      And I am in a 1913 building with old copper. Absolutely no way will I get a T-1. Did you forget that part too?

      1. Curmudgeon

        Have you looked at colocating a bare server?

        You’d need to hire someone to manage it and keep WP updated, but owning the hardware gives you more resiliance against abusive business models.

      2. tawal

        Hi Yves, I think that you should inquire about the pricing of hosting at WP VIP. Since it appears that you are going to stay with WP, they may provide the best oveall experience. Good luck and thanks for all that you do, tawal

        1. Yves Smith Post author

          You missed the comment in my post.

          WP VIP has dropped serving customers of my size. Their minimum per month is $3750. That’s absolutely nuts at my traffic level.

          1. tawal

            Oh, OK, now I see it. Yes I agree, that’s prohibitive and likely outside of your business parameters. Just occurred to me that WP is probably like all rentiers, drug dealers: Give it to ’em free, cheap … until they’re hooked.

      3. Nathanael

        If you can hire your own sysadmin, you’d be set.

        That’s pricey, but you probably know people who would be able to do it and would do a ood job, if you can afford them.

  5. Kristina

    Hi, Yves. Is it possible that your site has outgrown the traditional blog platform? Have you considered a content management system instead? I am thinking Expression Engine or Drupal (after a little googling, I found out that Gawker uses CMS for its blog).
    I am sorry about how WP treated you. I’m not a fan or a customer. That is just crappy.

    1. Yves Smith Post author

      HuffPo and FDL both run WP, with considerably more in the way of content, authors, and traffic than I have, so I am way way below the point where WP should be a problem, beyond the usual database issues. This is a pretty simple site. I am not changing horses. I don’t have time and a big change is an enormous amount of hassle, plus a lot of cost too.

      1. Brooklin Bridge

        I imagine that HuffPo would have their own servers and IT staff including a code-head or two for tweaking. Either that, or they would have a dedicated 3rd party service that was functionally the same thing. They make big money, no matter to which side of the hall Arianna is currently sending her clients.

        You can have very good software that still requires maintenance though here it sounds like reasonable software with a crappy, or at least lopsided, company offering the service.

        1. Brooklin Bridge

          Since WordPress is open source, it would be easy for HuffPo or FDL to have at least one developer on hand who could integrate, add functionality, and chase down bugs.

  6. Middle Seaman

    Most of what people say above is correct. IT is problematic, organizations are disorganized, some you can do yourself, etc.

    Most economists have no clue, most IT people are good only for simple tasks, most physicians can deal with hypertesion if it isn’t too complex and the like and nothing else. Managing medicine (e.g. office, labs, scheduling) is by and large done as if it’s 1950. The reasons doesn’t matter here.

    Advise: simplify everything, we read NC for its content and excellent writing, never buy complex deals, today you can split your service: one provider will do backup, another will do security and yet another will do daily operations. This approach is typically reasonable priced and doable.

  7. Finnucane

    The crumminess of our age. I drive down the main drag of my exurban-style town, and I see churches, fried chicken restaurants, pawn shops, liquor stores, check-cashing places, title pawn places, “bad credit/no credit” used car lots, video poker dens at gas stations – the detritus of a decaying empire. In such a howling wasteland, one should not expect honesty from a salesman. Or from a politician or banker, either.

  8. OMF

    The better IT shops are the ones run by actually techies. And the best ones are those run by techies, for techie minded people who are essentially just left to their own devices.

    You’re infinitely better off getting a general purpose server from a hosting company and installing wordpress yourself.

  9. Fraud Guy- Also

    Don’t dismiss the role of venture capitalists (in this case, Silverton Partners) and other outside, self-described “angel” investors in WP Engine. They are, in business-speak, “driving for an exit,” which can often mean sacrificing the long-term, when they won’t be around, in favor of the short term, which is when they will cash out.

    1. McMike

      Driving to the exit, control fraud, canibalizing… enriching yourself by destroying businesses; it’s the only business model these days.

    2. LifelongLib

      All this is the result of companies being run in the interest of absentee investors who don’t care about anything except rate of return. It’s been happening for a long time. Back in the 1950s engineers at Ford would beg for money to develop new cars, and the bean counters would tell them it’s cheaper to just put fins on the models they already have. It started with the biggest companies, now it’s everywhere.

  10. geof gray

    Devolution=American service by monopolies and quasi monopolies, e.g. comcast, TBTF banks, etc etc
    What Yes describes is the Big Grab in the wild west where internet hustlers are racing to fill a market space. Companies of the big grab are immature, developing. The trope is wrong even though the service quality is similar.

  11. Bert Markgraf

    I’d give Pair Networks at a call. My company has provided IT support for small business and professionals since 1994 and we run about 100 small websites. Pair has been our supplier for top quality hosting for the past 15 years. We’ve looked at their WordPress offering but haven’t needed it. They say they will install it for you and manage it. They have 24/7 tech support which has been excellent when we’ve needed it. We’ve found that the only way to navigate IT is with personal contacts and long term experience with particular suppliers.

    1. Yves Smith Post author

      Hah, I use Pair for Aurora, have been with them for 16 years, maybe longer.

      First tried putting NC on Pair years ago, immediately ran into DB problems and had to go to another host.

      You are missing the issue. The problem is NOT the host. I have no complaints about my current host.

      The problem is WP. I had wanted a WP specialist because they charge a premium to bundle in WP SOFTWARE support. I can either pay for the WP support separately and have much less than 24/7 or have more like 24/7 if I go to a WP specialist. Having one person (as someone suggested below, which is what I have now) does not get me to 24/7.

    2. quixote

      I can second the good words re Pair. Been with them over ten years for three small sites. Their tech support is instant and competent and has to be experienced to be believed. They’re also not cheap.

      But Yves knows all that. I’ve got nothing in the way of useful suggestions. It looks like the field of keeping medium-sized sites up and running — not just wordpress and not just thinking sites, but also small businesses — that whole field seems to be a magnet for the dregs of the tech world. Probably on the assumption that this is a ready-made pool of know-nothings who can be safely fleeced since they’re unlikely to sue.

      As an earlier commenter said, this is why we have regulatory frameworks. If we had a government instead of a kleptocracy, we’d have had something for the internet ecosystem when it first went commercial.

  12. Eric M

    I started working on a port of WordPress to Rails. It’s kind of a beast. Perhaps I will start it back up again.

  13. sd

    My impression is no one wants to work any more. They just want to go to meetings, be paid for their ‘ideas’ and collect big checks.

    There are pockets of people who work hard – but the people who collect big checks do their best to keep them hidden.

  14. John Moore

    I recently left a job as a military contractor. The Army was paying for premium support from a major OS vendor. My colleagues would submit tickets and watch them languish. I and my friends would often find a workaround or fix before the vendor’s tech support just by using some creative search queries via Google. I suggested to the Army that we use a similar product that wouldn’t cost the Army a dime because we certainly weren’t getting the support we were paying for. One can argue that the vendor will keep any important fixes within their own database away from Google’s clutches, but the vendor’s community forums are usually indexed by Google and up-to-date solutions are posted there. I wish I could give you better advice, but really good technical troubleshooters are few and far between.

    1. Massinissa

      Im assuming the military was using that particularly expensive service because either

      A. Someone making the decision to buy that product for the military gets a kickback


      B. Someone who produces that expensive product has high connections.

      Or hell, it could be both!

    1. Mannwich

      Rackspace is who we use, and we pay basically based on how much traffic we get. Seems to work really well for us.

  15. Bob Morris (@polizeros)

    PS The developer of a very large blog migrated it to Drupal a while back because he thinks WordPress at root is badly written and that there are way too many updates.

    I’m not suggesting Drupal (he said the migration was a bear) just that a seasoned developer saw major flaws in WP.

  16. Brooklin Bridge

    I’ve never played with WordPress, but my experience with Open Source projects has been that they tend to be a little clunky (they grow and grow) and the groups that develop them become sort of insular and tend to look upon the collective result with a sort of “hallowed” reverence as if it were more than just software. Outsiders who find the modules difficult to fit (integrate) into their own packages are the profane, to be tolerated only when necessary, etc., etc. They can still be excellent packagers as long as one is willing to do things in the cananocal way (which often means completely revising one’s own software to fit their model).

    I don’t know how much this fits WordPress, and since NC is not using any customized version, it wouldn’t really be particularly applicable, but issues solely related to the software would tend not lend themselves so much to the notion of devolution as to an inherent weakness of the development model. That and the fact that databases, particularly small ones, are notorious for needing maintenance.

  17. Brian

    For someone that has been through this as a small business owner, your best bet is to take all the bull by the horns. Do not parcel out parts of the job. Get one person to run the site. (they should be able to run several and monitor them without stress) If you choose to have a designer, then be sure they will hand over all of their work to your monitor. Pay for a hosting service that only hosts and has reputable up time, or better, do it yourself. But the one best thing that kept it all going was to host the server. The more control your controller has, the easier their job will be, which is to make it work all the time.
    This sounds simplistic, I know. But the fewer parts, the less that can go wrong.

  18. AndyLynn

    wanted to second the Pair.Com recommendation. i used them for *years* before down-sizing to their PairLite.Com offering; both services have been very very good/stable.

  19. bob

    It’s the app.

    It’s becoming pretty clear that all of the ‘new’ apps are just a front. In back there are thousands of PEOPLE running around trying to make these things work.

    But…I don’t want to deal with PEOPLE, I just want a button.

    It reminds me of the old gentry houses where the staff were kept in hidden tunnels. How did dinner get on the table? It just appeared!!! Magic…there’s an app for that.

    1. McMike

      Well, the executives and investors may want to act like software just happens with a magic button, but we customers just want to have the salesperson’s promises kept, not get lied to, and to talk to people who have the knowledge and authority to fix things when they break.

      1. bob

        “Should be” and “is” are two very different concepts.

        These days you sell people on “should be” and give them “is”.

        Don’t get me wrong, I appreciate the frustration, and yes, it’s most of the time the result of very bad/dishonest sales.

        1. McMike

          Strikes me that a high proportion of the people visiting this site would like to inhabit the land of “is” for a change.

  20. Pal

    I am a 52 yr old techie. I have programmed professionally for over 27 years. I am self taught in every language I ever got paid for… I learned Cobol in MIS and never used it professionally. A lot of what you all are saying is in fact true about IT. I don’t want to waste too much of your time.. so I will get to the point. I just retired from programming and I worked the last four years in Silicon Valley for Ebay, IBM, and ecommerce startups… and I retired because outside of Silicon Valley there has been no increase in real pay for most of ten years. I get offered contracts every week to program in web projects for the same rate in worked for in 2003! Are you kidding me.. to do ecommerce dev. you must be expert to very good in five distinct, interdependent programming languages just to do web sites. I do about 8 languages myself, some major others minor but all necessary. Five languages to keep up with and constant change and no real industry pay raises in 10 years…and you wonder why no kids want to go into it and old fckers like myself are quitting? I switched over to PM and Biz Analysis which I can do in my sleep in get paid 4/5 of the crap pay as a programmer… no stress, no languages, no nights and no weekends working. The industry gets what it deserves as the CEO and big talkers pocket all the money. America has screwed itself so many ways no one wants to bend over for fear of getting it in the wrong way. We are a nation whos time has passed. I see an America too stupid for me to live in. I have started padding my nest and officially moving to South America in Jan to watch bikinis the rest of my life. AMerica is not worth the price anymore; better places to live and much cheaper.

  21. McMike

    Reminds me of a company president I worked for. After getting sold one too many pieces of vaporware that landed us in court over non-performance, he said the next time we buy software, the sale presentation is going to be videotaped and made a part of the contract.

  22. Wangston Philberry

    “and please don’t tell me the issue is the hosting, not the software. The very existence of WP-focused hosts who charge a hefty premium to standard big iron hosting should be sufficient proof that this is a real issue”

    Hmm. The issue is actually that you’re too cheap to pay for the services you need.

    There is no free, open source, out of the box software that scales easily for all sites. It would be very difficult, if not impossible, to build such a thing. So you’ve got two choices if you want to build a site that handles any kind of traffic. 1. Write it yourself (expensive). 2. Use something like WP and then contract with someone else to scale it (expensive).

    Like many people, you seem suffer from the fantasy that technology problems are all easy to solve, so if someone isn’t solving them for you for free, or very cheap, you’re convinced you’re being ripped off by evil parasites.

    It’s not that easy. I’m not saying the WP Engine doesn’t suck. They suck. But you’ve got an expensive (for you – it actually doesn’t sound like your traffic/situation is *that* difficult to solve) problem, and the only way to solve it is to spend money investigating a variety of vendors until you find one that can solve the problem and provide good service.

    Like any outsourced service, you need to have a backup plan in case your vendor fails (expensive!) and you have to know enough about the problem to make an intelligent choice of vendor (c.f. Munger).

    Since there is a dearth of compentent WP hosts in the market, you might want to look at independent contractors and small web development shops as well.

    1. Yves Smith Post author

      Sorry, Blogger was just fine. It was stable and had decent functionality. Software does not have to suck. WP sucks. But Google is shutting down Blogger.

      This is basically arguing for a lousy platform and saying the problem with it are the users. No, Linux and NeXTStep and later (not to the same degree) OS X show operating systems do not have to suck. Blogger showed blog software does not have to suck.

      And as for “you need to pay more” I happen to run a business that won’t support more. And in case you missed it, the service is actually mainly OK, even if I’m having occasional annoyances.

      Oh, and you’re “suck it up and pay WP Engine” after they’ve lied to me repeatedly is not the solution. And your surmise, “their price is right price” is also unwarranted. Another guy who is well regarded and is about to start hosting WP sites as a speciality (he already hosts some, just hasn’t made it a focus) was going to do the migration the second time I tried with WPE and remarked, “no way is her site a $XXX site.”

      1. Wangston Philberry

        Oh no, I don’t want you to suck it up and pay WP Engine – I agree that they are terrible. I just think that you have to suck it up and pay *someone*.

        You mention Blogger as proof that free blogging software that scales okay is possible. Then you note that Google shut down Blogger. Hmm! I suspect itcost Google quite a bit of money to create and run Blogger, and they shut it down because it wasn’t making enough money for them to be worth the investment. It was never free – Google expected to make money off your site (and everyone else’s), and when they failed they shut down the service.

        WordPress is a different proposition. It’s actually free – dollars and speech – so when you need more than it provides out of the box it’s up to you to pay for it.

        I don’t know enough about this site, or WP Engine’s prices, to know if there’s a way for you to make things work at a price that you can bear. If not, you should definitely consider other platforms and services as well!

  23. MaroonBulldog

    Before you sign up for any IT service, peruse the agreement and make sure you understand what it promises you. Most IT vendor paper will be very clear concerning your obligations to the vendor, but the vendor’s obligations to you may be lettle but vague happy talk.

    Pay espeical attention to the service levels–do they warrant explicit times to repond to inquiries, restore service, and resolve the problem? Do you get a stipulated remedy if they fail? Or do they just name service levels for minerals, like silver, gold, and platinum? If so, ther minerals you are likely to experience may seem more like coprolites, plastic, and mud.

  24. washunate

    Yves, this is a great write-up.

    Perhaps the angle I would try to add something to the discussion is to observe that this is what happens as we have squeezed wages over the past few decades. It just isn’t profitable to provide high quality professional services to individuals and small businesses – the small customers can’t afford to pay for the service. The companies selling the service know that, so they low-ball and cut corners, then the customer is surprised that it costs more than the original low-ball to do the job ‘right’.

    You see it in firms that do have some pricing power, like established law firms. You get billed .1 hours at $250 per hour just to send them an email or have a 2 minute phone conversation. Or HVAC – try having somebody out from a reputable company to do work on a rooftop unit. There’ll be a trip charge, truck charge, or minimum billed time even if they’re only onsite for half an hour. Psychiatrists routinely charge three figure hourly rates. And so forth – all without meaningful guarantees that it will actually be ‘quality’ work that solves your problem.

    1. MaroonBulldog

      In my experience, software developers, systems integrators, and consultants treat their big business customers just the same as they treat their small business customers: they low-ball bids, cut corners, complain they can’t make any money at the settled price and need more to do the job right, etc. Whenever they can, they will write a contract that promises only an effort, not a result. They will charge by the hour, rather than by the deliverable. They will say they are offereing warranties and insert contract terms that do the opposite, limit and disclaim warranties. Oh, and their projects will probably take at least three times as long to complete as they promised. It sounds ironic to refer this industry’s offering as a “professional” service: it’s about as “professional” as a financial institution that specializes in making unsecured loans to unemployed persons for the purpose of financing down payments on used cars.

  25. Banger

    Unless you have to, avoid a redesign. I suggest keeping you sight and creating an experimental site if you want complex searches, taxonomy, exotic metadata, AI, and multi-media and in that case just get a couple of the hippest people you know in the biz to take it on or ask a team of your fans to come up with some plans in exchange for fame–have a competition maybe. And try to get away from WP in the long term.

  26. Ted

    For a site like yours, WordPress is probably the best option. WordPress is crap, but so is everything else out there. There is not a magical option, and WP actually works pretty well.

    WP isn’t unique or particularly shitty. WordPress is actually rather streamlined, particularly when compared to software packages like Drupal. If a page request takes 10 database hits in WordPress, it will likely take 50 in Drupal.

    Scaling sites is hard. Any site that gets dozens or hundreds of comments per article, and posts multiple times every day, is going to be a challenge.

    You are correct that your challenge is the database, but the answer is in high end hardware — SAS hard drives, high levels of RAID — and perhaps a caching strategy. These are not cheap.

    I hear you on the desire for affordable, managed, high-end hosting. My company is trying to offer just that (for Drupal not WP, sorry) and it is a serious challenge.

    For every $100 clients pay us for managed hosting, over $40 goes to a data center. After that, we have the usual overhead: rent, utilities, taxes, service providers we utilize, etc.. Once these are paid, we have to pay employees to write and maitain the code that manages the sites, test that backups and updates are working properly, investigate any issues that pop up, communicate with clients, and more. Then we still need to perform invoicing and payroll and whatnot. Which leaves almost no profits.

    We have also run into a weird problem: people don’t trust us because we are too cheap. At least twice so far this year, I have been told by a client that their website is mission critical and they need to pay more for hosting. Not get more, but pay more! Our prices make us not credible for managed hosting to some…

    And they might have a point. If we had a bunch of employees with salaries and benefits, our price structure might not be sustainable.

    We don’t have sales people, which has worked so far but won’t if we want to scale up. The primary reason why my company doesn’t have sales people is that sales people have consistently got us in trouble by overpromising. So, I hear ya!

    I don’t have any answers here. Affordable, performant, managed hosting is often promised but rarely delivered. And I kinda understand why: it’s hard.

    1. Brooklin Bridge

      […]people don’t trust us because we are too cheap.
      Frequently, that excuse is one of the aliases for, “not big enough”. You can have the best, cheapest package in the world that solves hunger and energy in one fell swoop and if you are not both big and established, you’ll get the most creative reasons that all boil down to just that. I don’t understand why they can’t just come out and say it.

      On another item you touched upon, I was under the impression that fail safe hardware has gotten extremely cheap over the last ten years. You can now get highly robust and highly configurable storage systems from places like Equallogic (now Dell) that can automagically mirror data or store simply the differences, hot-swap disks, auto-manage load, real-time diagnostics, and so on for as little as 5-10 thousand dollars for many (many?) terabytes of storage. One such system could easily handle several accounts like NC. No? I readily admit I may be missing something.

  27. aidee

    I’ll flag MediaTemple as a great hosting provider and have been with them for 5+ years. They would be regarded as a premium service but pricing isn’t dissimilar to the many out there.

    Currently use both a GS and DV solution, the later I suspect would be fine for NC:

    Re support they have been very responsive considering their location on the US west coast and me in the antipodes.

    Have looked at other options but cannot find reason to switch.

    Happy bill paying user and no other affiliations :)

  28. rotter

    This isn’t just a problem with software or support services its everywhere. Most consumer products are junk. “Profits” are made by driving wages and compensation down/firings & layoffs and avoiding taxes, to cheapening the quality of materials and design. It doesn’t really matter what the product is. OEM factories in the far east that produce this landfill stuffing have as their primary companies the investment/capital management groups that import and market this garbsage..the end user is not even a consideration (let the American worry about that)…and we shouldn’t marginalize this as mere “consumerism” or feel guilty about being pissed off..because underneath it all is the criminal capitalism that assassinated and canabalized the mighty American manufacturing economy that created the middle class and afforded the masses some measure of economic and political power…The architects of this disaster are perfectly, accutely aware of this. Economic and political power among the working class is not anything they approve of…and so we arrive at where we are today. Paying 3 or 4 times as much for miserable quality garbage products than we paid for much higher quality products when we made them…and/or unemployed,nearly politically disenfranchised – or in the case of the millions warehoused in our for-profit-private run prisons, totally and utterly disenfranchised.

  29. Brooklin Bridge

    This seems to me like an opportunity for a small development company with some cash.

    Devolution certainly comes into play. Speaking in the context of developers and development IT (source control, build teams, QA, etc.), things sort of fell off a cliff after the mini crash of 2000-2001, though the momentum for the crash had been building all through the Clinton years. In the 90’s It took a long long time for most of the mid to large size software companies to figure out how to do anything at all with their off shore development investments. For ages they spent fortunes bringing in whole cities of engineers of all skill levels over here, training them, and then shipping them back to lead the way, but usually that ended badly in the early days since an equally perplexing problem was local corruption and lack of resources in the foreign locale. But they kept at it, starting with the deep-pocket companies such as HP, and the bemused smiles of the in-house developers didn’t last long. Then they had to integrate with existing in-house code (sometimes monster code packages) and that usually meant changing their in-house high talent for more inexpensive and inexperienced developers (lots of students) willing to do grunt work under -occasionally- really nasty super high pressure, sweat shop conditions. Not an easy transition for company or developers, or documentation, or anyone, which gave a lot of rather empty satisfaction to the high-test engineers as they vacated their offices and cubes in wave after wave of wrenching lay-offs, but didn’t do anything for their stability of employment. A huge number of them were swallowed up by start-ups, but then most of those went belly-up somewhere between 2000 and 2009. A lot of them found themselves suddenly and rather magically to be 55 and older and simply had to get out; there was no easy path and often no path at all to a decent paying job. Systems programmers, language designers/writers, those specialized in other domains such as math, or law or finance, with a degree in CS and other really highly qualified engineers fared considerably better, but even they have taken a hit in the last ten years as foreign counterparts started to become available for that level of work at significantly reduced costs.

    And the task of running a multi-national software development concern may well have reached certain hard limits in terms of what can and can’t be done. It remains a logistical nightmare and the end result is frequently, but by no means always, of inferior quality to what would be accomplished by in-house local development from start to finish.

    Sales people in software were pretty much always flakes, amazingly even in some of the really big reputable companies. I remember some absolutely hair raising stories of the lengths of duplicity (staging heart attacks and such) sales people would go to for the sale. And the marketing division would generally best be left for some circle and sphere of Dante’s Inferno. So while it would seem logical that devolution would hit sales and marketing sooner or later, it would be hard to pin down when the old fashioned circus clowns, but highly knowledgeable circus clowns, and the evil marketing groups that could strip the soul out of even the most useful, brilliant software package, morphed into the un-careing and underpaid shills of devolution. Moreover, really excellent mid-sized software companies such as Oracle and Digital, had their ups and downs which may have looked like devolution but was really just a case of bad management.

    But that’s the software industry in general. Open Source Packages such as WP are slightly different and the companies offering hosting services specific to WP and other content management systems are also different. Open source projects almost always have problems that are mostly due to the model of develoment and not to devolution per se. And the specialized hosting companies have always been in a very high competition sphere and they have always been mercantile and mercinary and narrow minded. I suspect it would be even harder than for the software companies to draw the line between the issues they or their counterparts of the time had five or ten years ago and now and pin point where things shifted from more classic negatives to those originating from devolution. I suspect it would have been just as frustrating ten years ago as now, but of course for different reasons.

    As to solutions, not that Yves has asked, I suspect no matter how painful, NC will have to: 1) change packages with a lengthy evaluation process conducted by a wizened Yves and Lambert in order to maintain the same level of resources and cost (assuming such a package and hosting service exist), or 2) go through some painful additional research into companies for hosting WP at current levels, or 3) and I don’t know the problem well enough to know if such would be appropriate, but anyway, get a consultant involved (a one time expense) that can track down the exact problem(s) and offer solutions, work-arounds, etc., within the context of which ever hosting /WP specialized company and server set-up NC decides to use.

    On another somewhat related subject, I wonder what opinions Yves, Lambert or Others have of the comment package, Discus. It seems to be gaining popularity. The Rolling Stones web site as well as TruthDig (Chris Hedges) have recently converted to it. Unfortunately for me, it requires JavaScript which I have turned off. I’m not sure why JavaScript is suddenly becoming popular again, but it seems to be for a number of things besides comment packages, yet I know it used to have a number of security issues and so I’m hesitant (though I imagine they have fixed that old problem at least until they need to fix it again). Still, I hear Discus is interested in making money out of (cornering the market for???) comment sections and I assume that also means data mining them. I wonder why someone like Chris Hedges or even Matt Taibbi would go along with such an obvious scheme.

    1. Brooklin Bridge

      What Discus appears to be doing strikes me as at the heart of devolution. It’s not simply a collapse of resources but also of integrity. Data mining people’s interests and points of view, their political, sexual, health, shopping habits and other characteristics, and not seeing any problem with making a profit off of that; now that is devolution.

      1. Brooklin Bridge

        I had not looked into Discus very much so your opinion is the first I’ve heard. I assumed you weren’t considering using it, rather I was curious as to what you thought of it. I assume then, that people like Matt Taibbi and Chris Hedges really have little or no say about the matter. Particularly Chris Hedges; the system seems pretty incongruous with his points of view.

  30. Cujo359

    This isn’t uncommon for WordPress. It apparently needlessly complicates the creation of dev sites or staging sites,

    Yes, for instance, and while it’s a simple thing this is something that makes it a whole lot easier to have a test site colocated with your real one – WP doesn’t honor port numbers in its host URLs. They deliberately remove them. I don’t know why.

    WP is OK for blog software, I think, but I’d hate to be dependent on it.

    I’m writing this without having read the preceding comments, so it’s possible that someone else has said this, but these days, I wouldn’t try to set up any site I had to depend on without being able to do a thorough dry run. That means build the site the way it should be, have people try to use it the way they would the real thing, and only when it passes all those hurdles would it go live. Anything short and you’re almost certain to fail.

    Of course, you can still fail even having done that, but at least it’s not certain. If you’re planning on starting or transitioning a big site, I’d be wary of anyone who says otherwise.

    Speaking of Lambert, his recent adventure upgrading his site are probably a case in point. Even as careful as he was, there was a period when the new site was down almost as much as it was up. It turns out Drupal has a few idiosyncracies, too. Go figure.

    I wish it weren’t so, but that’s the way things work in this big libertarian paradise of Internet services.

    1. Brooklin Bridge

      I wouldn’t try to set up any site I had to depend on without being able to do a thorough dry run.

      That’s an excellent point and the ability to do so is not simply critical for a new installation, but also for debugging situations that arise out of the blue long after an installation has had the kinks ironed out.

      1. Cujo359

        Not to mention complex systems that depend on lots of software, which tend to require an almost nonstop stream of patches to stay reliable and difficult to crack. Back when I was working with an outfit that was using commercial software platforms, every upgrade was a crapshoot. You never knew what would stop working when you applied a patch. Now that I’m mostly using FOSS I’m a bit more blase’, but that experience still makes me nervous when I apply them to a system I’ve come to depend on.

  31. Brian Daly

    While I certainly empathize with the plight described, I don’t see how this can be described as some kind of devolution in business.

    WP is free software,that offers a huge amount of functionality. The where does the idea that it should scale and be relatively trouble free to host a mid-traffic blog for dirt cheap come from? How was this possible in the past? Maybe Blogger, but that came from the relatively new world of ad supported software, and Google seems to be getting serious about supporting the services that pull their weight in revenue generation.

    Basically, the whole product category (off the shelf CMS) is relatively brand new, less than 10 years old, so there hasn’t been time for industry maturity, let alone devolution.

    As many of the commenters have alluded to, the technology space of website hosting is undergoing a state of rapid, constant change.

    FWIW,if the costs of running your blog which reaches millions of people are less than $10k per month, historically that strikes me as an astounding reduction in cost (compared to pre web days).

    Figuring out how to support the cost of “content” creation has always been a tough nut to crack.

    I built a site for my business on Drupal, and it too has the issue of everything getting stored in the database, with the attendant problems you’ve mentioned.

    1. fajensen

      The Common Problem with CMS-systems is the use of MySQL ;-)

      Depending on the configuration (MyISAM or InnoDB), basic SQL functions like referential integrity and type checks has to be done either manually in code (MyISAM) or not (InnoDB).

      The CMS developers cannot assume that they can control the database configuration so they have to code the core for the worst case, MyISAM, thus adding bloat and inefficiency.

      Other developers, f.ex. people writing modules, do not always bother with defensive coding because MySQL “runs” the SQL anyway – MySQL will just skip instructions that are not supported by the underlying engine, usually leaving data that should have been deleted in the database.

      Then there are the quirky parameters that influence performance. A common problem is that when the index tables grow large enogh, there is a huge performance hit when the tables cannot fit in memory. If replication is used together with lots of writes, then performance will also suffer. e.t.c.

      Balancing all of this is difficult. The actual problems depends on the situation, even the honest and competent CMS-provider will have a hard time optimising a particular site properly. Normally the cheap hosting places keep costs low by using use the same Virtual Machine image for every customer and the bare minimum of configuration afterwards.

      In the end, someone, who knows what they are doing, will have to tweak the site configuration. This work costs money either up-front with the generic hosting or by extortionate monthly “service” fees from a hosting provider (who might not be competent nor actually doing the work they are paid for).

      And – that is why many people use generic hosting.

    2. Yves Smith Post author

      Thanks but you are overestimating our traffic. 1.4-1.5 million PAGE VIEWS a month, about 700,000 unique vistors (which is unique per IP address, so a person using a work computer, a home computer, his cell and his iPad to view the site would count as 4 people).

      1. Brooklin Bridge

        There are consultants, particularly database consultants, who can often convince a vendor to let them look at their hardware/software on behalf of a client (NC) so as to get to the bottom of some issue. Moreover, they would typically know up front (before such an arrangement) whether or not it would be useful -solve or mitigate the issue once and for all) or if there is simply some problem with WP and/or it’s use of MySQL, or simply with MySQL, that can’t be addressed at all, or that requires the ongoing watchful eye and occasional tweaking of an IT person, or some such that may or may not be acceptable to your needs.

        1. Brooklin Bridge

          Of course you may already know the broad answers to these questions and not need a consultant to tell you what is obvious.

  32. MarkJ

    Allow me to add my two cents worth and pass on some of my web site building and moving experiences – limited as they are.

    Fajensen is spot on with the WP hosting server analysis. The main criteria for a web hosting provider is that they provide data in a timely matter to the web user, the site may be easily ported into the hosting provider’s backend software and allows the administrator to easily maintain the site.

    I will say that IMHO WP has enough flexibility to perform basic blogging functions but lacks a large base of third party application (add on module) development. One of the open source packages that offered better third party module development is a package called Joomla. Joomla uses the same back end as WP (MySQL, PHP and Apache). Unfortunately, Joomla’s most recent incarnation did not have a third party comment module at the time the most recent version was released. That omission made Joomla useless as a blogging platform. There may now be a comment module for Joomla that does support some type of blogging or commenting function (that was available in the prior version of Joomla).

    The provider that was eventually chosen to host the website was Hostgator. The provider seemed to have enough server hardware and a large enough internet connection to sufficiently handle the user’s web presence requirements. One of the criteria that was used when selecting the provider was the apparent lack of negative comments about the provider in the blogsphere (a technique that is often used). The hosting provider also provides the MySQL PHP and Apache virtual server components that both WP and Joomla use. Also, they have good technical support and even made an unsolicited call back concerning a support issue.

    Eventually, someone will have to be responsible to make sure all of the hosting requirements of the provider are met, is able to port the site to the provider and is able to administer the site. It is at this point costs escalate if the site owner is unable to provide these administrative functions.

    Disclaimer: I am not paid to provide recommendations for any product or service mentioned in this post.

  33. Prometheus

    Hi Yves.

    Excellent post. It’s hard to argue with your conclusions. But there is a glaring omission here I feel compelled to point out. Your problem is pretty simple. You’re a financial analyst. You are trying to promote your work with the Internet. The internet is a big, complicated machine, rather like your car and the road. Now what you want is for the car and road to “just exist” without problem or complaint. That’s a reasonable thing, but here’s the catch. YOU DON’T WANT TO PAY FOR IT. Yes, there I said it. You have to hire a professional to maintain that car, and pay taxes to maintain that road. What you are doing instead is basically being your own mechanic and road crew. This works for many small operations ( if I only drive around the town and take occasional road trips ) but not for you, you are now too large to do that. You must ( holding my breath here ) CREATE A JOB. Yes, you, dear Yves. Check that bank account you’ve got. See all those zeros at the end? Now, here comes the barrage of numbers and calculations to show just how poor you are, how little the site makes, and how it just doesn’t make sense to have a dedicated person. After all, THAT IS YOUR JOB. And as I’ve surmised from reading your blog over the past few years, you’re probably pretty good at it. I’m sure you’ve improved the bottom line for many of your clients. But that is what has brought us to this point. People are simply refusing to pay for goods and services, yet expecting ever greater amounts of both, and are shocked and astounded at the results. The hostile responses, bait and switch, and all the rest flow from that basic problem.

    Now I know you’re a smart cookie, and if I wasn’t pointing a finger at you, we’d just agree and commiserate over the state of things. But I’m looking to be a little provocative today, in hopes of maybe shedding some light on _why_ this real and serious problem has absorbed us all. Things really are falling apart, it’s not just an illusion.

    At the end of the day, you must pay the Devil his due. It’s not much at all, really. Why is this so hard for everyone? If we can answer that question, I can show you a world you people can only dream about today. It is only this one step away.

  34. Alec Kinnear

    Hi Yves,

    We run a WordPress migration service (mainly Typepad to WordPress but we moved The Hollywood Reporter weblogs from Typepad to WordPress and then from WordPress to Drupal) and have just the by invitation only (who the heck wants to deal with walk-in traffic from spammers and scammers of every time or employ a large sales team which forces a growth binge and compromise of quality and core values) hosting service you are looking for.

    Very speedy dedicated servers (dedicated to you and not the other clients sharing some giant cloud) and VIP service at about a fifth the prices. We’re also developers of both commercial plugins (FV Flowplayer 5) and popular open source plugins (FV Simpler SEO, FV Antispam Foliopress WYSIWYG).

    As migration specialists with an SEO background, moving a busy site like from one host to another is as ordinary as eating breakfast for us.

    Our clients for both development and hosting include sites like Professor Juan Cole’s Informed Comment and Ray Make’s busy fitness site DCrainmaker who are roughly in your traffic weight category depending on current events. We also host some very busy client celebrity sites who are about four or five times as busy as NC but I won’t name them here. Feel free to contact either of them (I expect you know Juan Cole already) for a reference.

    Other attribute, you’d like: we’ve been in business for 13 years, formally as Foliovision for six years. We have real physical offices and the people who would work on your site have worked together for many years. As the posters above noted, it’s hard to hold a tech team together, but we do our best.

    Outside of hosting, our comment management plugin “Thoughtful Comments” would help you stay clear of rubbish like Disqus by giving you advanced comment/er management inside WordPress (Thoughtful Comments makes it much easier to deal with large quantities of comments and especially to deal with difficult political crowds and the consequent vandals).

    For some reason Thoughtful Comments never took off and Disqus did and I’ve never had enough time to do something about it.

    Amusingly enough, I gave Naked Capitalism to three people for Christmas this year. Never thought I’d be bidding on the web hosting contract.

    I hope you do manage to solve the issues! If you’ve got good hosting now, feel free to give us a shout if you need some custom development.

Comments are closed.