The miserable business of making a living from web analytics software

In another interesting post on his blog, Avinash questions the conventional wisdom about “Enterprise Class” software, specifically web analytics software. The post made me smile because of Avinash’s hyper-cautious preamble about how he doesn’t set out to be contentious, but sometimes he just has to speak his mind, yada yada yada. I had a mental picture of him donning helmet, mitts, shin guards and chest protectors (not to mention the other thing) before continuing with the rest of his post.

But in the end Avinash has not had to endure a hailstorm of foul tips (see how I worked that in? You’d almost believe I had the faintest idea what I was talking about); his comments box is full of people agreeing with him that “Enterprise class” is a bogus term usually used by software companies to provide an excuse for charging a lot for software.

But Avinash’s post did make me think about the bind that companies who want to charge a lot of money (or even any money at all) for web analytics are in – it’s actually pretty difficult to build a business on the back of this kind of software.

Violins, please

If you’re trying to build a business on the back of paid web analytics software, you have a problem. And that problem is that, however busy the E-metrics summit may get, and how many thousands of people read my blog on a daily basis (that’s about 0.1 thousand, since you ask), web analytics remains a niche activity. This limits the total amount of money a web analytics vendor can get out of a specific customer.

Many of the worlds biggest software companies (and Microsoft is probably the prime example) have got there because they have been able to cut deals with customers to put their software on hundreds or thousands of desktops across the business. The numbers quickly multiply: a $100 desktop app (or OS)sold for 10,000 seats nets you a million dollars – and that’s without implementation, server-side stuff, upgrades etc.

But this is never going to happen with web analytics, because the user base within the customer is just too small. I might go to (say) Coca Cola Corp and sell them a fancy web analytics implementation for maybe a million bucks, but once I’m done, what can I sell them next? Well, the first thing I do is make sure that I can charge them a million dollars next year as well – this, more than anything else, has driven the web analytics industry to an ASP model. And hopefully my company’s product range has something else in it that I can sell into the customer this year. And, of course, I can sell service.

Service!

Service is great because it’s the gift that keeps on giving (to the vendor, that is). Compared to going and finding a new customer, or even persuading an existing customer to buy something new, extending a service contract is easy peasy. But the problem with service is that it’s low margin. Before you know it, the software company that you thought you ran has turned into a professional services company; your developers end up spending time out on the road with customers because it’s hard to find and train consultants quickly enough, and it’s even harder to schedule their time efficiently.

Before long, this reliance on service for revenues can have a detrimental effect on the product itself, causing it to become harder to use and thus more reliant on service. What it certainly does is provide a poor incentive for a software vendor to provide a solution that can get up and running quickly with the minimum help. It also provides a poor incentive to vendors to walk away from deals which will “run and run”. When was the last time you heard a rep for a software vendor say, “you know, if we implement our software in your environment, it’ll take ages to get value from it, and you’ll end up spending four times as much on the implementation as you budgeted, so we we’re not going to bid on this deal”?

So “Enterprise” web analytics companies face a number of options in order to keep growing – find more customers (difficult as the top end of market starts to really solidify), build or acquire new things to sell to their existing customers, find new ways of extracting more money from existing customers (sometimes referred to by the technical term, “price gouging”), or find other ways of earning money (for example, by using ad revenues to subsidize product development).

This is why you get such heartburn (as a paid web analytics customer) when your traffic goes up and so does your web analytics bill. It also explains why there’s a slide in every Omniture presentation which is headed “How we will take over every aspect of your business”. And, of course, it’s why, if you buy web analytics software, you can expect to have a very close relationship with your account manager.

Disclaimer/please don’t hit me

You might think that my position here at Microsoft makes me biased against paid vendors, and that it’s easy for me to take a pop when we don’t have to charge for Gatineau. I can’t change your mind about that, but let me just say that I have a great deal of respect for all web analytics tools out there, and think that a paid model is an essential part of the industry. My remarks in this post are borne from six long years of trying to make web analytics pay in an Enterprise vendor. It’s hard work.