How Audience Segmentation Increases Revenue for Publishers — Talk with Madhura Sengupta
This week on the podcast we talk with Madhura Sengupta, who is the Director of Ad Product Technology at Edmunds.com. Madhura has been a frequent speaker and contributor to the ad ops world, so I was excited to have her join us on the podcast to talk about audience segmentation.
Segmenting an audience allows a publisher to provide better targeting for advertisers and, therefore, command higher rates. Edmunds has set up their own data warehouse that is used to track, store, and segment their audience in extreme detail. This detail allows them to offer major advertisers, such as Ford, the ability to influence potential customers in the crucial moments before they purchase a car.
In this podcast, Madhura walks us through the logical and technology that Edmunds used to build this process. Audience segmentation would be helpful for any publishers, and especially those with audiences interested in major purchases and high value goods.
1:26 — Madhura’s background
2:17 — How long have you been in the programmatic space?
3:00 — What is the general theory of audience segmentation and why should a site do it?
3:40 — How does Edmunds segment their audience?
5:25 — Do you target based on cookies or 1st party login data?
7:00 — What are the types of things that Edmunds tracks
7:25 — How do you track that this data leads to purchases?
8:24 — What technology are you using to track your audience and build this system?
9:35 — How long have you been segmenting data like this? What do you do with it?
11:20 — How do brands use this data?
12:35 — Do you get a sense of what difference audience segmentation make in revenue?
13:50 — What is the main question to ask for audience segmentation?
14:33 — Do you get the sense that Edmunds is more advanced at audience segmentation?
16:00 — Does segmenting your audience lead to less valuable segments?
17:23 — How does Edmunds use this for look-alike modeling
18:42 — Is this data most valuable being sold directly or is it used programmatically?
19:45 — Would the data you have be valuable to other people in the industry?
21:00 — How would a site set this up for the first time?
22:22 — Technically speaking, what tools do you need to segment your audience?
23:35 — How evolved are brands in purchasing like this?
24:35 — Do all car brands buy like this?
25:52 — Message to the larger ad ops world?