Avoid early split testing

Conventional wisdom says that you should split test your adverts. Write at least 2 adverts for each ad group and show one to the first searcher, the second to the next etc and watch the click through rates till you get a clear winner.

It’s a great concept and makes perfect sense. After all, who wouldn’t want the best performing adverts they could get.

There are a couple of practical problems with this though:

  • It means writing at least double the number of adverts. Coming up with one set of great adverts is hard. Coming up with a second set of great adverts is even harder.
  • It can take a long time before each advert has been shown often enough to get a result you can trust.
  • The most clicked advert might not result in the most sales. Scouting the trail a client took all the way from an advert to an offline sale can be like trying to track a cat across a highway*. You can’t ask every client “Did you click on the advert that said ‘Great deals on air conditioners’ or the one that said ‘Great prices on air conditioners’”.
A more significant problem is that focussing on improving the first step in the sales process distracts you from doing other stuff that would have a bigger affect on your bottom line.

Let me explain.

One of the things we noticed in the our earliest AdWords efforts was that one client would rate the leads we found for them as great while another client wouldn't make a single sale out of very similar enquiries.

The first time it happened was with a legal firm. We were finding people who wanted ante-nuptial contracts (prenup agreements) for them. They didn't make a single sale out of the 25 enquiries in the first month. They cancelled.

We sent the next month's enquiries to a different firm who closed. They made five sales for every six contacts.

The sales funnel in both of these firms was pretty much the same. Here's how it worked:

  1. The happy (but cautious) bride or groom would search Google for an ante-nuptial contract and see our advert.
  2. Some of the people who saw the advert clicked and ended up on the specialist lead generation micro-site we set up.
  3. Some of those visitors would then fill in the contact form and end up in the lawyer's inbox.
  4. The lawyer would call to gather information and prepare a sales quotation outlining the service and the price.
  5. Some of these potential would accept the proposal and the work would be done.

The two firms charged almost the same price for that service, but there was one significant difference: the first opted for efficiency and scheduled all the phone calls in a batch at the end of the week while the second called the prospective client as soon as they possibly could.

No amount of split testing adverts would have improved the results the first firm got. Their sales process was broken.

If you’ve got a steady stream of enquiries from your online advertising campaign it’s far better to improve your sales process first and then go back and look at improving the click through rate. Optimise from the bottom of the sales funnel first.

*Disclaimer: no cats were harmed in testing this theory.

© Peter Bowen 2018 | Isle of Wight