I was recently invited to present at Ken McCarthy's System Intensive 2-day seminar in London (November 2008). I delivered a presentation on "Web Design 2.0" - addressing the relationship between Web2.0 and graphic design.
While there was heaps of excellent content and insight, the biggest thing I took away from the seminar was the vital importance of experimenting, testing and comparing all marketing messages, not just putting something out there and hoping it works OK.
The great Drayton Bird, one of the most successful direct marketing copywriters ever, followed me, and presented some fascinating evidence of DM messages (both in mail and email), asking the audience whether they would expect item A or B to have performed better. Half the time, the expert audience got it wrong.
This throws out some interesting points. One is that common sense won't always tell you accurately whether your message (web site/brochure/mailout/newsletter) will work well, or even whether one will work better than the next. The other key takeaway is that the only way to find out is to test! Professional marketers have been doing this for years, but how often do web designers bother to test even two alternative designs?
I'm came away determined to find evidence that saving pixels actually works! I know it's common sense, I can provide logical proofs why reducing noise and focusing attention on indicators to the path forward should increase success rates, but can I prove it?
I'm setting out to find case studies from real-world web sites that have implemented a "Save the Pixel" type approach, in such a way that we can isolate the factors, with web stats that prove a change in success rates (whether page views, increased conversion rate, or some other success factor).
So here's some interesting early evidence, taken from this site.
A week ago, I implemented a few changes to the Save the Pixel site. I removed the solid graphical header, and replaced it with a larger heading1 (which will be different on each page, so reducing "dead pixels"). I also rewrote some of the intro copy, and added clearer navigation in the left hand column.
Of course, the copy changes and addition of navigation aren't stricly just pixel-saving, so this isn't hard & fast evidence, but it's interesting nonetheless.
Here are the early results, from Google Analytics, which show a marked increase in goal conversoin rate. The goal in question is simply people proceeding to the purchase page. Normally, it has been around the 10% mark. In the last week, it has jumped by approximately 50% to almost 15%.
Interesting for sure... I personally suspect that adding the permanent side navigation will have had a significant impact, likewise the introductory text. Simple but relevant.
How much of a difference will removing the top bar and adding a more impacting H1 have made? The only ways to tell for sure would have been either:
- to make these changes in sequence, a few weeks apart,
- or to do split tests, showing a proportion of visitors one design, and another an alternative, to get more accurate comparative data.
But that would have been more work, wouldn't it? Maybe that's why we (I) don't get to the bottom of these facts, to find out for real whether these assumptions are quantifiably provable. When you have a good hunch, a combination of logic, common sense, and style, you can get away without the extra work.
Well, that's not enough for me any more. I'm looking for real evidence. And I'd love your help. Here are a few ideas:
- I've implemented simple split-test functionality in my CMS to help make it easier to test alternative design solutions.
- If you've recently redesigned your site (preferably changing only the design, not the content), I'd love to find out what difference you've seen.
- I'll be looking to do the same on any future projects we get in at Scratchmedia.
