I’ve had a great day at the web analytics congress, the keynotes by Neil Mason and my good friend Avinash Kaushik were inspiring as ever, and Mieke from Independer.nl had a nice presentation. Not much real news in there for me, but very, very inspiring.
What I saw today was mostly this though: even though we marketers know how to talk in KPI’s and metrics, and we think we know how to set these, most of our clients don’t, yet. While we talk about whether we’re working on the completely correct data, and how we should attribute parts of a sale to different channels in complex sales situations, most of our (future) clients, would be benefitted more by just starting to measure those sales, and change them from page views to some raw estimate of income.
I’ve heard weird stories today, f.i. Mieke from Independer openly proclaiming that they attributed every sale to it’s last referrer. Even if that last referrer was an obvious navigational search. Now I can say it’s wrong, but… it’s working for them. And even though I think that Independer should know better, as one of the few companies I know in the Netherlands with their own 5 man strong in-house analysis team, I’m wondering if I’m not being overly arrogant.
If you’re a geek, like I am, and love to play with numbers, like I do, you probably have a tendency of making analytics overly complex, of over-analysing. And that’s bad.
Why that’s bad? Well, because Mieke can explain what she does and how it works to her management, and I, who would probably try to give a percentage of the sale to the search and a percentage to the earlier referrer, would probably have a hard time explaining it to anyone not in our industry…
If you’re able to run a profitable business by keeping it that simple, and simply accept that some of the data you’re measuring isn’t completely correct, aren’t you being a lot more efficient?
I’d like to say yes, but in the end, I think it will work against you. Independer works in the insurance business. I think it’s safe to assume that people shop around quite a bit before they get a new insurance, going through all the stages of the AIDA process. That means that quite a few people might come to your site from early on in that process. Them seeing your site and your brand is very important at that point.
If you, however, cut costs rigorously, and throw out every campaign that isn’t profitable for you, like Independer said they did, you will throw out all of the sites that help people find your site early on in the orientation process. If you stop that campaign, you will see your ROI go up. But will you also see the missed sales? No you won’t. You’ll be fooling yourself with analytics.
So it might work for them now, but at some point, there’ll be a competitor who’s better at it, and they’ll beat them at what they think is “their own game”. Making analytics too simple is a risk, whatever excuse you have.