Content pfp
Content
@
0 reply
0 recast
0 reaction

Dan | Icebreaker pfp
Dan | Icebreaker
@web3pm
Still surprised how often companies mess up customer segmentation by exclusively using behavioral clustering The most stable and intuitive segmentations rely on demographic clusterings. These still often have strong behavioral correlations, but the demographics are usually the **cause** of the behaviors, as well as **differing responses to the same treatment**. The last point is key, since without demographic segmentation, you won't learn that treatment 1 (e.g., 30% off) works best for segment A while treatment 2 (e.g., vip upgrade) works best for segment B. This is an ancient talk but the principles (8:30-10:00) still apply https://youtu.be/g-mRWavmXuk?si=LFsadSaUyltPBTNq&t=506
2 replies
1 recast
7 reactions

Apurv pfp
Apurv
@apurvkaushal
we typically do segmentation after having spoken to / observed 10 or 100s of users - that helps us understand which markers matter in terms of segmentation.
1 reply
0 recast
1 reaction

Shira Stember pfp
Shira Stember
@shira
Great insight on segmentation blindspots! Who your users are (demographics) often matters more than what they're doing (behavior). This is why getting to know your users is so important. Knowing what drives them, what frustrates them, etc...matters so much when building anything. Sure, all those heatmaps and UX tests give you valuable data, but they're missing the bigger picture of the actual people using your product.
0 reply
0 recast
1 reaction