Pattern #31: Instant Results - Increased Leads For Vivareal.com.br
Not all patterns will have a strong (+10%, +20% or +30%) effect and that's good to know. We observed such a small effect from using instant filters that would show results instantly after each filter change (instead of requiring an additional search button click).
We tested this little change recently with Vivareal.com.br, a real-estate site in Brasil in one of our mentorships. It was Rodrigo's idea to remove the button and update search results instantly after any filter change. Afterall, it's a popular belief that making people wait less and providing them with results sooner might be a good thing. And so here is the actual A/B test which explored instant filters:
This change boosted Leads by 1.6% with a sizeable margin of error (+/-2.1%). That's not much at all, but the fact that it might have a small effect means it's worth testing further. So, we turned this test into a pattern.
This is how most of our conversion patterns start. At first it's just a single test. But then we collect other data to see if the effects repeats across different types of sites and industries.
Here's Pattern 31: Instant Results
- Instant Update
Search results should update instantly after each change to a filter. Not much to it.
Let's Discuss This Pattern
What are some situations that might pose a challenge to this pattern?
What if search results cannot be shown quickly due to elaborate calculations? Is that an indication that the calculations themselves should be simplified?
How do we deal with text input fields? It may be feasible to update the results in real time, as the user types. But what if the process is expensive for the back end and confusing to the user? After all, it may be distracting to see all the interim matches as I type. Still, lots of sites do it well, like Netflix.
Should we have a small submit button next to text input filters? Should inputs show Autocomplete suggestions underneath, which is less distracting because it's plain text and we are all familiar with it.
We'd love to hear your suggestions. And if you want to test this pattern or adapt it to your needs, we want to help you run it as a test along with many other improvements backed by data as well.