I had the opportunity to interview Ms. Chen Zhao about data analytics in the online advertising business. Ms. Chen Zhao is a leading expert in data analytics for ad targeting using behavioral data. She has over 16 years of experience working on various strategic and customer marketing challenges addressed through the analysis of digital advertising data. She has developed and refined her analytic framework and models having worked at leading companies like Monitor Group, Razorfish, Microsoft Central Marketing and Amazon advertising. She holds MBA from IMD, MS Mathematics from Dartmouth, and BS in Mathematics from Denison University. Aside from building very powerful predictive models for customer acquisition, she is a mother of 2 boys and lives in Seattle, WA
Click on the link below to listen to the interview. If you prefer to read the transcript, read below.
Ms. Chen Zhao Interview on data analytics for online advertising
1) What has changed for online analytics for advertising over the last 3 years?
Three things have changed:
I. The multitude of channels and therefore many data sources. But there is still the same customers, only that we have more touchpoints to reach them. Therefore we need to integrate the data across platform and channels to have a single view.
II. The response metrics have also exploded. At the beginning it was just the click, and then we all look for the ultimate action – a sale, a registration. Now we have “engagement”, which can mean anything from interaction with a rich ad unit to video viewing, to scanning QR code.
III. The blurring of the dividing line between online and offline.
2) What is still the same? How should a brand manager think about the explosion of data from digital advertising? What should they pay attention to?
The idea of reaching the right customers at the right time with the right message is the same from the age of direct marketing and TV. The concept of customer segmentation and targeting is still the same (driven by basic economic theories – the supply demand curve).
The explosion of digital data can be neatly organized into two categories:
I) The stimuli (or the drivers) from the multitude of channels – usually taking the form of ad impressions, and all the meta data that describes these impressions
II) The responses – what actions customers took as a result of viewing the ads. Different brand managers would be interested in different metrics, but they all boil down to the conversion funnel. Basically you can neatly line up all the response metrics (or the ad effect) on a continuum from awareness, interest, and consideration to purchase/conversion or even repeat.
Analyzing the data boils down to relating the stimuli to responses and identify what types of stimuli drive the best responses (i.e. what combinations of ads best drive your objective metrics
The most important thing for brand managers to pay attention to is clearly articulating what your objectives are and what your success metrics are for any given digital advertising. We are often fascinated by all the technology and all the fancy things you can do with the new channels. But at the end of the day, if you cannot measure success and don’t know the return on investment, you will not have what you need to convince your CMO to invest more in a program. Once the objectives are set, all elements of a campaign – targeting, creative, channel selection, media weight, promo, etc – should all be aligned with the objectives.
3) Lets look at the problem of customer acquisition. What type of data from advertising should brand managers/media buyers pay attention to improve customer acquisition?
Identifying the right target is critical in acquisition. The right targeting can bring tremendous efficiency for advertisers in terms of $, and at the mean time provide the best customer experience (relevance). The data that is useful to create the best targeting is not necessarily limited to demographics. Behavioral targeting is an increasingly important area. Targeting should leverage the knowledge about customers’ interest/intent based on their past browsing, searching, shopping, purchasing behavior. Contextual targeting is also important. Again, relevance relevance relevance. The more digital channels we have and the more noise there is out there, the more important it is to help customers cut through the clutter.