I thought I would reference and summarise a lot of articles out there to solicit and continue the debate of a website redesign for and against conversion rate optimisation. What is better and why?
The answer is purely contextual but the below articles should help and solve the debate. In my opinion, conversion rate optimisation is more effective than a website redesign for multiple reasons, but the overriding being that a website redesign, based on ‘good UX’ and solid insights or not, changes too many variables at once, meaning that we don’t know which variable affects the users experience positively or negatively. However, a website redesign sometimes cannot be helped due to legacy, technology or politics (the latter of which can be persuaded otherwise in some cases)
Did you know that 68% of marketers did a site redesign in the last 12 months and 1/3 were unhappy with the results? Rich goes on to say, “And you really don’t want to waste an average of $55,000 dollars on a failed website redesign do you?”… he’s got a point! He continues to provide 7 core reasons as to why this is the case from too many opinions, the process itself taking too long, too little feedback but my favourite being that ’too much is changed at once’ echoing both Paul and myself above. In fact, he nicely summarises it with this graphic.
Paul echos the above stating that redesigns change thousands of unknown variables at once and comments that many decisions are based on intuition or who shouts the loudest. In turn he recommends, not really to adopt a process of conversion rate optimisation as in some or most instances it is not a substitute, but instead to focus on a data-driven website redesign, a good recommendation being to A/B test your design ideas.
Following on from the above thoughts, Paul focussed on the M&S website famous case study (where they resulted in an 8.1% drop in sales). Again, he spoke about a DDR (data driven redesign) and minimising risk of the redesign by controlled experiments and insight. I think this screenshot best explains his thought process
Peep sums it up in a sentence with “The most typical scenario is that there will be no difference in terms of performance. The reason is that some things get better, some get worse – and they cancel each other out.” showing examples like Marks and Spencer and Digg. Whilst he continues to say that optimisation is a much better route, he, like I, agree that there are times when a radical design is better, including: tried lots of optimisation already, technology being severely outdated, design of the site is amateur and / or there’s very little traffic. I would personally add a few more to that list and ever question the 3rd point made.
David Mannheim, Quora post: Conversion Rate Optimisation is often preached as better than redesigning a website from scratch. What is your experience of redesigning a website from scratch when faced with the choice of CRO or a website redesign and how did you choose?
This sparked a really interesting debate from just a total of 5 answers. Jesse Bilsten talks about technical debt - an interesting debate where introducing a feature costs us more than half of what is would cost if we redesigned due to the current technology solution. David Somerfleck continues by stating that “Competent web design (to me) can't be separated from internet marketing. So it's either a whole unit done right the first time, or you scrap it and try again.” - weeding out the free (or in my words, “shit”) websites from those that perform. Finally, David Levine believe that CRO work has a lifespan before a new design or architecture - that being of about “5 - 10 years worth of work”).
A somewhat impactful headline (nice work, Chris!), but discusses the points well, focusing more on an ‘evolutionary site redesign’ (ESR) than the dramatic alternative. Why? Well he states that within the 2-5 years that your website is being redesign or not, it’s staying ‘still’ whilst the rest of the web is improving in what he calls “the wedge of suck-ness”. It’s not that it’s bad, “…but a better, and less risky, approach involves a process of testing with incremental (and often dramatic) improvements”. He goes on to give the reasons why of which have been mentioned already.
This was an interesting article preaching the use of AB testing tools (specifically, and somewhat in a bias fashion of Optimizely) for a website redesign. The example used is Formstack.com where learnings from AB tests helped inform website design decisions within the design process.
Don’t forget that radical changes to your site, especially it’s structure, can dramatically affect SEO, too. Therefore I would recommend reading an article such as Richard Foulkes “A step by step guide to updating your website without destroying your SEO”
Bonus: Did I forget to mention that Users hate change ?
Overall, I echo my original thoughts at the beginning of this debate. There are much more risks with a site redesign than the alternative of continual optimisation. However, perhaps there are alternative options here than ‘just’ CRO and a website redesign’. As Paul mentions above, data-driven redesigns do and can work, especially those in my opinion that are more evolutionary in scale that radical - something which Peep argues for using the case of Amazon and how ‘similar’ it looks now to what it did in 2005.