And on the 8th day, G-d made clickstream analysis and it was awesome!
Up until recently (about 2 years ago or so) online marketers had to rely on was the last click before conversion when tracking results.
To illustrate what I mean by this this lets pretend you’re Sir Phillip Green and just before taking your Friday afternoon private jet to Monaco (as you do every Friday), you decide to look at your Adwords campaign and see how your keywords are doing. You notice that on the keyword “womans fashion” you are getting 1000 clicks a day at a cost of £1 per click. This keyword also returns you only 10 sales a day. Now you decide to look at your keyword “Topshop” and you notice whilst you only got 100 clicks they only cost you £0.05 a click and you made 20 sales.
So if you were a traditional old school type of marketer you would say au revoir to those nasty generic terms and bonjour to more brand terms.
Right?
Wrong. So any online marketer worth his salt will tell you. Why? Because anyone who has ever bought anything online will tell you that most purchases start with broad generic searches. It is only over time and through exposure to several brands and products that the decision is taken to buy. And once this happens either the brand or product is put in the search bar and the sale is made. So to look at the last click is to exclude yourself from the consideration set that one needs to enter at the beginning of a search.
This has been touted as common sense, the only problem up to now has been putting a cost on what terms drive your brand into the consideration set. Knowing how likely a keyword will lead to a purchase 3 or 4 clicks down the line is a key determinant in working out how much I am prepared to pay for it.
The technology exists to do this either through Dart or Atlas, but the quantitative analysis needed to drill down into the mass of search data, makes this sort of insight available only to those with the budget to pay for it.
I saw a presentation today on some clickstream analysis that was done for one of our clients. It was so beautiful it made me want to cry. We have some serious quant jocks at the agency I work at and they trully did what is easily the most complete piece of clickstream analysis I have ever seen.
The analysis showed that for every online sale, there was a research time of 2 weeks on average in which the average sale took place after 4 clicks. (and the majority of last clicks being on brand) By knowing this and what keywords were used at different stages of the buy cycle, we were able to adjust our bidding and our messaging to target buyers with critical messaging earlier on in the buy cycle. The change we saw was immediate. The average clicks per sale dropped to 2 and the amount of sales on generic terms went through the roof.
Such a clever piece of analysis coupled with actions that drove value to the client…… *sigh* its the stuff sweet and profitable dreams are made of
Since Google bought Doubleclick we have all been expecting to see additional features being rolled out within Google Analytics that would give us this kind of insight. It makes sense from a Google point of view for advertisers to be able to see the value of any given click instead of the spray and pray approach we all have to our paid search generic campaigns.
The oil of this industry is numbers. Hard, cold, honest numbers that set this type of marketing apart from any traditional techniques. Clickstream analysis, or at least the lack of it, is one of the few final frontiers of uncertainty with regard to the true effectiveness of any online campaign. We need to know how all the pieces of our digital jigsaw fit together in order to assign value to our spend. Once things like IPTV become more mainstream and as mobile phone handset technology begins to allow richer ad formats begin to be served, knowing the true value of how the media meshes together is going to be fundamental to taking this industry to the next level.
