Experiment gets almost no traffic because another experience outranks it
Some numbers first. Home hero experiment launched 8 days ago: 40 visitors total, 3 conversions. The same page gets roughly 60k visits a week, so the experiment is seeing about 0.05% of eligible traffic.
We also run a broad geo-based experience on the same slot, and it has a higher priority than the experiment's experience. My working hypothesis is that nearly everyone matches the geo experience first and never reaches the experiment.
Is that how priority is supposed to interact with experiments? And if so, what is the intended pattern here, reorder priorities or restructure the audiences
1 answer
That is exactly how priority works. Each experience has a unique numeric priority rank per slot, and when a visitor matches multiple experiences on the same slot, the higher priority wins. A broad geo experience ranked above your experiment's experience will absorb nearly everyone, which is why you are seeing 40 visitors instead of thousands.
Two ways to fix it:
- Reorder the priorities in the dashboard so the experiment's experience outranks the geo one. Visitors who match both will then enter the experiment.
- Narrow the geo experience's audience so the two stop overlapping, if you want both to keep running with their current ranks.
One more thing worth flagging given your numbers: starvation also stalls the statistics. Every variant needs at least 1000 visitors and 25 conversions, plus a week of runtime, before the experiment leaves the "in progress" state. At 40 visitors in 8 days it would effectively never resolve, so fixing the priority is what unblocks the stats too.