Good coverage starts with deciding what matters. Not everything gets tested equally. Nor should it.
The first step is to figure out which of your user journeys are non-negotiable. By non-negotiable I mean, they must be tested. And, if it's not obvious, the tests must pass before you ship.
I don't have an exact science for how to do this. I usually start by asking what would break the business if it failed.
For an e-commerce site, checkout is non-negotiable. I lose money whenever a screen reader user or keyboard user can't complete a purchase. For a bank, not being able to have access to your account and make payments is a catastrophe.
Now I have a list. Then the work begins.
I test screen reader compatibility on all those critical paths. I make sure keyboard users can complete those core tasks. Only now I'm no longer guessing. My tests cover what people actually use.
When I'm done with critical business failures, I can ask which features would cause the most frustration. A decorative carousel breaking is annoying. The search function breaking is quite bad. And these are the next tests I do.
Here are some examples.
- For an e-commerce product, I focus accessibility testing on product browsing, filtering and checkout. I can live with the blog section having lower coverage because users with disabilities rarely visit it.
- For a project management app, I prioritise task creation and assignment because that's where users spend the majority of their time. I accept that reporting gets less rigorous testing.
- For a marketing site, articles, media and navigation have priority.
But wait, you ask.
What happens with the stuff you don' test? Do you just ignore it?
Not quite. I document it. I know where the gaps are and I'm transparent about it. I don't pretend 60% coverage is 100%. I don't need to. I know these features I've put aside for now are less important and I mark them as lower priority and come back to them as time allows.
I know there are gaps. But I am intentional about them. I've made a deliberate choice. I haven't ignored them completely.
Big difference.