How not to make money

2012-11-19

People that know me know I don't really like to work at companies where income is the main driving force. If your owners and/or managers are purely driven by income (ROI) all the programmer will be doing is trying to fit in another 30x180 rectangle ad in an already overcrowded website. Not fun.

And if you're anything like me, you too will despise the mindfucking/tracking that marketeers try to do with visitors of a site. Yugh.

Oooon, the other hand. I do realize that companies need to make money. I need to be paid, just like anybody else. It's just that I think there are also other ways of upping your ROI without treating your visitors like cattle, or worse, sell them to the slaughter house (-> advertisement agencies) en masse. These involve making a better service, more tools, fun (premium) stuff for users to play with; a better user experience in general. If you give a bare but good user experience for free users and promise (and deliver) them an even better experience for the premium versions; people will pay. And they will pay much.

The problem with making any kind of platform is not making the site. That's a problem that can be easily fixed and something you can control tightly. The biggest problem is getting a userbase. Users are precious because you can't really force them to join your service, regardless how awesome it might be (for them, or in general). They have to come to you somehow. This is the biggest paint for a start-up, and under appreciated by vested sites.

Let's jump to the case in point. Before putting my blogs here, I used to have a weblog at qfox.web-log.nl. It's Dutch, but don't worry, you won't care about it anyways. It's mainly highly opinionated stuff related to philosophy, politics, news, or random shit. Definitely my pre-JS university era ;)

So web-log ("web dash log") used to be a pretty popular blogging platform in the Netherlands. As most popular services, they started out simple and became progressively more complicated and cluttered. As they became more popular the inevitable point came where they had to introduce ads to individual blogs. It's the inevitable turn-around point for the popularity of any service. As soon as you introduce ads into your site, you'll have entered the all-or-nothing phase. Like many I fled the scene immediately and started on my own blog (here ;).

I never deleted that blog though. Never touched it again either. I'm fine with the content remaining online (and actually haven't been confronted with any of that content over the course of time, so far, for whatever that means). At some point, while I was still blogging there, the service joined hands with another service (TypePad I think? I believe they hosted the editor and some/most of the content). Never liked that step myself, but that matter didn't then nor does it matter here.

About two years ago, I think, the service was bought by a Dutch newspaper company. I didn't follow it closely but at some point the "TypePad" service pretty abrubtly (it seems) ended their collaboration. It took web-log over a year or so to migrate all the hosted blogs back to their own server. Most of them went pretty quickly, but a bunch did not. I read this in the news btw, so I don't know the nitty-gritty. Don't think mine was affected, but hadn't checked until I logged in today and it survived the migration, one way or the other.

Fast-forward to today. I received an email from the service that my blog was about to be shut down within one month unless I'd login. Say wut??

Now let me explain that. I was not surprised/shocked/appalled by the simple fact that they were taking my content offline. Listen; it's their service. I've never paid for it. I've not touched it in ages. For all they know I'm dead. I get it, and I'm fine with that.

However, I don't understand why a blogging service would deliberately throw away an (undoubtedly huge chunk) of their (presumably zombie) userbase and, more importantly, their content. Especially not under the pretence of "those old blogs bogging down the database". WHAT? That is your reason? *wow* Mind blown...

Now, I've worked at WaarBenJij (take a look, count the number of ROI driving bullshit above the fold, cry, and close), which is a Dutch travel blog. At the time there were eh, 2 million blogs I think? It was once, like web-log, a very popular platform. The number of users/blogs was something quoted very often, with proud. And I recall being appalled by that myself, because I of course knew that most of those 2M users hadn't logged into the service for years. So it felt like a cheap shot to me. But I did get it; You have the content, you have the (zombie) accounts, so you're allowed to brag with them. They can drive traffic. They can keep users on your site, driving up your ROI.

At that time, the master database was still running on a single server (32g ram or something). In fact it was just a simple master-slave where the slave was a replication server accessible to us in read-only mode, so we could run heavy JOIN queries on it or do backups without bogging the site down in any way. This worked great! I'm fine with telling this now because I know the server configuration has changed dramatically the past few years. I'm just trying to make a point, anyways (below ;).

Another thing I've learned is that while most accounts were zombies, occasionally some accounts would become regular accounts again (making those zombies more like comatose, I guess). In fact, some would also become premium paying customers in some occasion. Hello, ROI.

So web-log deleting old blogs and potentially comatose users goes against everything I've learned in the past years. And I wouldn't have been upset by it, if they didn't give us a stupid reason for doing so; site load overhead.

And what exactly should be the cost of running those extra blogs? Hardly any, on top of the already existing architecture you need for your current blogs. It's:
- a bit more bandwidth (but that's good, because traffic adds to the ROI, or should)
- disk space (mainly database text, cheap)
- migration pains (ok, this depends on the way you develop, but you'll HAVE to migrate blogs anyways, one or many should really not matter)
- exhausting the pool of short usernames

We're still talking old blogs here, so it's mainly text, right? Text is cheaaaaap. And if it isn't just text and media content is really causing a pain for you, you could simply change your policy such that abandoned blogs are only viewable in a slimmed down version (so that it's only text you need to serve, with ads of course).

Database overhead? Bullshit. You move the zombie blogs to a weaker database that's heavily optimized for read-only TEXT tables. You only migrate them back to the main server when they are re-activated. Won't bog down your existing server, won't make you delete huge areas of content. Won't make you do something stupid like lowering ROI.

So in short; I'm not angry because they want to delete my precious content (everything is exported to some word-press-able xml format, which is nice). I'm angry because they're giving me a very weak stupid reason for wanting to do so.

Somebody at web-log should be fired. Seriously. And I'm not even questioning the year+ migration track here.