When you’re losing 7 out of 10 people in the dark, you don’t wait around for a service engineer to fix the streetlights. You take a search party out with a torch.

3 Little Lessons in UX-Led Conversion

I was talking with an old colleague about lessons we find we have to re-learn every so often. Important things that we know, but that can easily slip out of the old brain-box.

Here are three I‘m sending to future-me so I get reminded in a year or so.

(In case you’re wondering what I’m on about, UX-Led Conversion is my term for making websites perform empirically better for businesses by making them better for the users.)

Lesson 1) Look where the light isn’t

A policeman sees a drunk man searching for something under a streetlight and asks what the drunk has lost. He says he lost his keys and they both look under the streetlight together. After a few minutes the policeman asks if he is sure he lost them here, and the drunk replies, no, and that he lost them in the park. The policeman asks why he is searching here, and the drunk replies, “this is where the light is.”

I’ve worked with several clients who were searching for their keys to better conversion rates under the streetlight.

One had quite a complex multi-step funnel. They were very data-driven and had plenty of data — although it was only on the later parts of their funnel. They felt good about this. After all, they were being data-driven, generating insights and testing hypotheses. That’s good, right?

They hadn’t even realised that they were searching under the streetlight. They hadn’t noticed this important fact:

Out of every ten visitors to the website, only one made it into the light; seven were getting lost in the dark.

Check your funnel. How many conversions are you losing in the dark? How many visitors leave without you having the slightest idea why?

Ouch, right?

It’s common that areas of a business get neglected. Often, they’re the areas where you can make a big impact.

So you need to start looking where the light isn’t. But how do you do that?

Lesson 2) User Research (because users are people too)

To understand why your visitors are vanishing in the dark, you’re going to have to find out what’s happening inside their heads. You’re going to have to remember that they’re individuals with feelings, opinions and fears; not a faceless mob of numbers and percentages trundling through your funnel.

Treat your customers like a faceless mob and sooner or later they’ll come after you with pitchforks and flaming torches.

And it’s so simple to do. Just invite some potential customers to meet you in person. Sit down with them. Listen to them. Empathise with them. Watch them (try to) use your website.

It’s so simple. But it’s not easy. You might be lucky and see some obvious easy wins, but it’s likely that you’ll have to work to distill the insights.

As Ben Holliday said on the Twitter, “user research is about finding the thing behind the thing.”

When you start to find the thing behind the thing, you will start to design better split tests. Split tests that will be more likely to make your customers happier and so much more likely to drive long-term gains for your business.

That’s when you need…

Lesson 3) Be careful what you wish for

Everyone who has run split tests has been disappointed by not getting to significance. It’s miserable when all your work setting up the test is in vain; when all your hopes as you start the test are slowly dashed on the cruel rocks of randomness.

It’s easy to compare your failures against the published big wins all over the web. It’s easy to forget publication bias: only the interesting tests that win get published.

You can start to believe that maybe there’s not much improvement possible. Maybe you just have to be happy with a 5% lift here, a 2% there…

I had one client who were confident their funnel was already pretty good. They thought they’d picked all the low-hanging fruit. They anticipated grinding out tiny incremental improvements.

Maybe they were right. (They weren’t.) But it doesn’t matter anyway.

The point is that if you wish for 5% lifts, then that’s what you’ll get.

Your subconscious brain tries to give you what you’re looking for. Meanwhile, it will actively filter out bigger ideas.

This leads to two problems:

  1. You overlook big, bold ideas that can get you the big wins
  2. You get bored of the small wins. (Hardly surprising. They aren’t much to get excited about — even if some do stagger over the line into statistical significance)

You don’t want this. You want your tests to win or lose unequivocally. You want to test big, bold ideas. So, instead of looking for small wins, try an exercise I adapted from Julia Cameron’s Morning Pages.

  1. Imagine yourself three months in the future, at the point where you’ve doubled the conversion rate. Yes, doubled. At least.
  2. Write:
How did we double our conversion rates in three months, you ask? That’s a great question. Let me tell you…

Then start writing freely. Don’t edit, don’t criticise, don’t stop. Let your subconscious come up with something. Do this every day for about 750 words or fifteen minutes. (I actually have a reminder in my phone to do this.)

Most of the ideas that pop out might be crazy. Might be terrible… Or brilliant.

Remember that you’re also loading up your subconscious with deep understanding of your customers through user research, so its odds are massively increased.

And you don’t need 50 brilliant ideas. You only need one.

  • An extraordinarily simple landing page that’s just this headline you wrote, an empty text field and a big orange button?
  • A super-long landing page that looks like a sales letter from 1927?
  • A video of your CEO explaining the idea to his 3-year-old?
  • Cat gifs?
  • Insert your ideas here.

Testing bold ideas can feel risky. But that’s only if you don’t understand risk. In fact, meek, “safe” tests are much riskier.

But what if the crazy idea kills conversion?

Simple: turn off the test. Split testing is a safety net that lets you try bold ideas and protect the downside.

Yes, there is a cost of a test idea that fails: effort to implement + a few lost conversions over a few days. But the hidden cost of not testing is much higher.

Your test idea bombing is not a bad thing. Here are two much worse things:

  1. No test ever gets to statistical significance in either direction. You’re not really learning anything. How long are you going to keep up the energy for testing when there are no results?
  2. You stop testing. Then you’re losing all the conversions you could be getting… forever.
“You miss 100% of the shots you don’t take” — Michael Jordan.

Woody’s Round Up

Remember it’s all simple, but none of it’s easy:

  1. Start looking where the light isn’t
  2. Start user research
  3. Start looking for bold test ideas

Do it now. Don’t put it off until you can “do it right.”

Because when you’re losing 7 out of 10 people in the dark, you don’t wait around for a service engineer to fix the streetlights. You take a search party out with a torch.