This past weekend I spent on Texel – a Dutch island – with over 200 experimentation and CRO experts at a conference called Conversion Hotel. Organiser Ton Wesseling saw me speak at uxcon Vienna 2024 and invited me to give an evening talk. It should be entertaining, of course, but also offer perspectives that are not that common in this group. So my dramatic passion for typography was a perfect fit.
A bit of a design outsider
What struck me was how isolated the experimentation niche felt – at least from where I was standing – from the broader UX and design industry. Or maybe I just didn’t see them before? Like their own little universe. Compared to UX folks, who might lean more idealistically, the experimentation crowd focuses on achieving and measuring real outcomes.
As an outsider to this industry, I learned a lot:
- The term “Experimentation”: I’d describe it like this – how can you craft scenarios and measure them (mostly with A/B tests) so you can make better business decisions?
- CRO (Conversion Rate Optimization) is a part of that, but not always the main goal.
- Growth hacking is not well received here. It’s seen as deceptive. This group wants to do things ethically, solidly, honestly.
- There’s usually limited design influence in this group. They work within rigid design systems – often stuck with mismatched fonts or styles set by marketing or brand agencies.
Typocalypse Now
In my talk Typocalypse Now I covered how bad typography can kill branding, usability, clarity, and overall experience. I think it became clear to many what the effects could be.
What remains open is – how can you measure good typography? When adjusting only sizes, weights, and spacing, the design looks clearer, more guided, more refined. It is obviously different. But does it make a difference?
Since this event, my head is spinning.
Let’s ask the experts
In an unconference session before my talk, I tried to pick attendees’ brains: how do you measure the impact of typography? There are plenty of studies that show how to improve readability, I know that. But how does this translate to actual websites?
Typography can be crucial for:
- Cognitive load – A fun term that came up: paragraph blindness. Like banner blindness, but for scary-looking walls of text.
- Form completion – Someone told me of a case where the label was too far away from the input field, and it tested poorly.
- Appeal – How trustworthy does it feel? How well does it fit the brand? These are trickier to test.
The usual basics still apply:
- Depends on user, context, and device
- Choose a typeface appropriate for the job
- Minimum font size: 16 px for copy, 14 px for UI
- Minimum text contrast: 7:1 for copy, 4.5:1 for other text
Setting a metric
First, you need to define a metric. By what will you measure the result? This part is tough and highly dependent on your product and goals. Potential metrics:
- Time on page
- Click-through rate
- Add to cart
There could be many more, I am not really familiar with this yet. Maybe you have more ideas?
What will you test?
We agreed it’s important to isolate typographic changes – separate from layout or other adjustments. Test them:
- Site-wide
- Limit to typographic changes only (sizes, spacing, weights)
- Alternatively, test font changes (same layout, different fonts)
This is a first step. I made a lot of great connections and I’m curious to find out how typographic changes perform out in the wild.
Thanks to everyone for the feedback, ideas on how to measure type, and all the great conversations! This was truly a transformative event with extremely open and friendly people. Even as an outsider I felt welcome. If you ever get the chance to go – do it. And for the US folks, there’s one in the States too.
What are your thoughts? Have you tried changing typography and measuring it?