3 ms·
OP here. > I'd love to see the actual numbers. The screenshot OP posted suggests they've been running A/B tests with 115 + 108 users, which is just not a big e
by gingerlime 8y ago
OP here.
> I'd love to see the actual numbers. The screenshot OP posted suggests they've been running A/B tests with 115 + 108 users, which is just not a big enough sample to derive any conclusions.
The screenshot is from the intercom.com website. Not related to our own experiment, which was run outside intercom.
I'm having trouble finding the actual raw experiment results, but I would imagine it covered 10s of thousands, possibly hundreds of thousands of participants, running for a month.
Regarding the other comments: yes, maybe we were doing it wrong. But we've given it way more attention and though than "hey, can I help you?" and really tried to target specific events etc. And also, we're a B2C so volumes are high but customer lifetime value is relatively low, which increases the acquisition costs.