The one where we redesigned the signup and messed up conversion
Written by Mario Ferrer & Sergio Vila
Season 1, Episode 23
FADE IN:
INTERIOR TECH COMPANY (OPEN OFFICE ARRANGEMENT) — MORNING
A product owner walks over to a product team, takes a deep breath and states
Product Owner: We have to redesign the signup! Look at Google, Facebook, and all those flashy apps out there that have millions of users. Our signup looks ugly and outdated.
Everyone nods in agreement. They use (and love) most of those flashy apps! So the whole team scrambles and starts working on a new signup. There’s a whole bunch of new functionalities to be designed, they spend months working hard to get everything just right, they launch it, and… BOOM! The new design performs waaaay worse than the previous one.
Product Owner: We have a 50% drop! This is massive. We have a problem. A big problem. A really big, big, big problem.
When something like this happens panic hits, and it hits hard. The product people start talking to each other in disbelief
Developer 1: But… we did everything right.
Designer 1: But… the new design was way more polished and functional than the old one.
Designer 2: But… <! — place excuse of choice here — >
Eventually, people start to get fired.
FADE OUT
DESIGN DIRECTOR’S NOTES
Test, test, and test again
The one thing that consistently causes drastic drops in conversion is a lack of testing. A lot of elements are updated in a new design: new functionalities, new flows, new copy, new visuals. But massive redesigns are generally not well received by people. Actually, they tend to hate them. If you launched the new signup to 100% of your user base without properly testing it, you may end up not only having one problem but two: the drop in conversions and not knowing what caused it.
Don’t do the obvious
The first reaction when a problem like this hits you is to roll back to the previous version. DON’T DO IT. I will say that again. DON’T DO IT. The temptation to go back to the solution that worked before the redesign can be very high, but in the long run, it’s worse to do that. Rolling back is not always as easy as it sounds (technically speaking), it will confuse users, and most likely it won’t solve the problem.
Yes, but…
Right now you’re thinking: Ok, I got your point but the problem is still there. Yes, I should have tested the sign up before launching it. I didn’t. Blame it on me. But the reality is we have a super-duper new signup that is making us lose people (and money). We need a plan to fix it now!
Ok, ok, here are some ideas that you could try to help you fix the conversion drop problem:
Test the task completion success rate
- Consistency in design is not as important for people as completing the task. You need to understand where they are dropping off to know what should be fixed. Data scientists and BPMs are your best friends here. You can also use affordable (even free) tools like Google Analytics, Crazyegg, or the fantastic Lookback to do live interactive testing.
- What if people just close the app without interacting with it at all? You won’t be able to understand these small nuances just by looking at the numbers. You need to get some qualitative feedback as well. You can do that by observing how people interact with the product. Bring some of them to the office (they don’t bite). DON’T over-explain, just ask them to perform specific tasks while you take notes. Believe it or not, if 5 out of 5 people can’t find the sign-up button, that’s a problem you have to fix. You don’t need to test with 2,000 individuals. You don’t need to have “statistical significance” to see you have a problem there. If sharing a common space with other humans is not your thing, you can also use session recording tools like Inspectlet.
- Take all the problems you have identified, prioritize, and follow an A/B testing approach to fix them. From there, continue testing, measuring, and optimizing.
Test the content
- This one may seem obvious, but even a product with well-written content and nice graphics can have a drop in conversion rate. Just because it’s good doesn’t mean it’s what people are expecting. Again, talk to them and compare your new design with the old one. You will discover what they understood in the previous design that they’re not understanding in the new one.
Understand people
- Yes, you should definitely have done this before spending 6 months working on a new design that is not working, but it’s never late to do things right. At the end of the day if you want your product to be successful you have to design it thinking about who will be using it. Not everyone has the same needs, behave in the same way or have the same motivation to do a specific task.
- A good way to know who you’re designing for is to create Personas. Yes, that thing everyone talks about but no one actually uses. Keep in mind that Personas are not just fake props based on fictional characters, you should build them by collecting real information from people, even from pre-existing knowledge in the organization. In any case, avoid building them based on assumptions or personal feelings. If you hear someone saying: “If I were a user I wouldn’t do that”, run away and set fire to that meeting room.
FINAL THOUGHTS
Following these steps and testing as you move forward will fix the problem. In future projects avoid massive redesigns and embrace subtle evolution (We recommend you to read this article from Jared Spool). When you get a better understanding of who the people using your product are, what are their needs, what they love, what they hate, and what specific problems they have; you’ll be a bit closer to delivering them a good experience. By constantly testing and iterating you’ll avoid pissing off people and not having 50% drops… and the best thing is, no one will get fired.
We recommend you to read the following articles. The authors are way smarter than us…
- The quiet death of the major re-launch by Jared Spool
- Three ways your new website will murder your conversion by Sayf Sharif
- Why your website redesign will fail to lift sales much by Rich Page
- Website redesign for higher conversions? tread lightly by Peep Laja
- Personas: why is it important to understand your users? by Keepitusable
Written by Mario Ferrer & Sergio Vila
Designfeld is a comedy series based on a typical tech company about a UX Designer and his fellow UX Writer, in their struggle to find… they don’t even know what they’re struggling to find… they just want to share their views on Design and why it can get quirky sometimes.
