<< by on October 23rd, 2012
Following my session today on retargeting was Chris Goward featuring the top ten conversion optimization mistakes. This is fresh stuff from Chris — he added many new points just last night!
Chris started with a sample from BabyAge.com. The goal was to lift ecommerce sales with their product page template. The test involved simply swapping the right and left columns, with image on left v. right. Variation B won with a 22.8% sales lift — which was surprising to the room because the image was on the left.
Now, the top ten mistakes!
1. Relying on “Expert” Advice without Testing
Chris showed how many “experts” claimed that a particular button color worked best. But Wider Funnel has found that many button colors work best — it depends on the context of the page.
2. Relying on User Testing
Many people think they are doing conversion testing when doing user testing. Often the Hawthorne Effect is in place — where they users are really being tested for the ability to try to perform well vs. a test for actual elements on a page.
3. Using Pre and Post Testing
This happens when you examine conversion rate with one page versus another one after the other, but not simultaneously. This doesn’t take things like inventory levels and seasonality into account.
4. Starting with Multivariate Testing
The myth is that you could put any variation into a tool and it will generate brilliant results. Often tool providers are focusing on complex tools like multivariate testing, but in reality Wider Funnel sees 90% of the results from A|B testing.
5. Using Software Vendors for Conversion Strategy
We assume that those that sell a tool know best how to get the best testing results out of the tool. But a tool vendor is best at creating robust, stable tools. Agencies are better at understanding your business issues because they focus on that every day.
For instance, Honda makes great cars, but would you have Honda teach your kids how to drive?
6. Measuring the Wrong Goals
Up front, Chris looks at the goals with clients. He called it the “goals waterfall”. Your overall business goals should flow down and define your marketing goals. Your marketing goals should flow down and drive your conversion optimization goals. There’s often confusion among these goals.
7. Not Developing a Strategy
There’s lots of talk about tips and tactics. Those can be misleading outside of a strategy. Think about the PIE Prioritization framework that Chris developed:
- Potential (heuristics, voice of the customer) – which pages have most potential for improvement? which pages are most important?
- Importance (traffic volume, cost)
- Ease (technical, political)
8. Not Using a Structured Testing Process
Chris shared a study from eConsultancy of companies that are doing conversion rate optimization. Those using a structured approach were realizing a 200% lift in sales. Chris shared the LIFT model:
Convert weaknesses in the LIFT into strengths. If the value proposition is missing “FREE”, for instance, revise the headline in the test.
9. Not Aiming for Marketing Insights
Chris shared an example from Electronic Arts. The goal was to get more game registrations — a key business driver. They wanted to do this specifically for the Sims3 game launch. Chris showed two versions of the page, version A showcasing the type of content you could get and variation B which featured gaining SimsPoints. Version B won. In another test, version A gave a free town if you registered your game vs. version B with the SimsPoints. In this case, Version A won, generating a marketing insight — for this audience, there were certain offers, even with reduced choice, could have a dramatic improvement. Now they are hiring more people to build items like the free towns to continue to make offers like this.
10. Giving Up After Inconclusive Results
If you have inconclusive results, don’t give up! Analyze WHY your tests aren’t performing, adjust, and test again. You can gain great marketing insights from this analysis — it can lead to massive improvement.