How to Apply Big Data Analytics in Retargeting
Enough and more has been written about the importance of big data. So we will not labor the point any further. These large, complex, disparate, and rapidly changing data sets offer the world of display advertising many advantages. Retargeting too utilizes big data in order to improve the experience for the consumer and also the results for the brand or the marketer.
What kinds of data are used in retargeting?
The two main types of data captured and mined for retargeting are user data and product data.
User data refers to the millions of users who visit the brand website and exhibit different forms of browsing behavior on it. Several aspects of user data are important for a retargeting campaign. As an example for an airline brand, user data will include:
The other important big data set is the brand’s product data. In several industries, especially online businesses, product data can be voluminous and change frequently. Imagine a multi-category online retailer and how rapidly elements like product prices, colour variants, and availability can change in such an environment. As a result, it is important for the retargeting engines to have accurate product information at any point of time. This may include:
How is this big data analyzed for retargeting campaigns?
The analysis of big data results in making retargeting campaigns effective for the user as well as the brand. Big data analytics aids processes like:
- Segmentation: Users who are most likely to buy and hence most valuable for the brand are identified from the others.
- Real-Time Bidding (RTB): Data is used to bid for the users’ impressions in real-time so that the brand who values the user most wins the impression.
- Product recommendations: Banner design and product placement on the banner is determined by a bunch of algorithms.
Every marketer knows all website visitors are not equally valuable as prospects. Some land on the website because a link is clicked by mistake or by chance, some others visit the website just for the information, and yet others are passing by. And then there are the loyal visitors, loyal buyers who enjoy spending time on the website and have a higher propensity to buy. Research seems to suggest that up to 75 – 80 percent of retargeting revenue can often be attributed to only about 15-20 percent of the retargeting impressions.
The variety of user data available helps brands create micro-segments and create individual retargeting strategies for each of them. As a result, brands are also able to change their investment strategies for each segment and maximize their marketing ROI. On the other side, users who belong to the not-so-interested segment are not bothered by too many retargeting ads just because they clicked a banner once.
Real-Time Bidding (RTB)
Here brands compete for the impression of the user and only the highest bidder gets to display an ad. The bid value is determined by the users’ data, the kind of product pages the user visited, and by usual buying trends on the brand website. As an example, the brand will bid highly for the impression a user who:
- visited in the last 3 days,
- spent more than 10 minutes on the website, and
- saw two of the brand’s highest-selling products.
Big data ensures that bidding is based on many more complex rules than the rather simplistic example above. Bidding ensures the brand who finds the user most valuable wins the right to utilize the impression. Also, for a consumer who is really in-market (the term “really” is used loosely here for a bunch of metrics that brands use to gauge the level of interest) for a particular product, banners are extremely contextual and hence useful.
Typically every banner offers the brand the ability to sell the product of interest as well as up or cross sell other products. If most users of an apparel brand website who buy a shirt are also known to buy ties on the website it would make sense to showcase both shirts as well as ties on the banner. Several such logical groupings are created by brands on the basis of click-through as well as purchase data analysis. Some examples include:
- Those who viewed product X also viewed products Y and Z.
- Most of those who viewed product X ended up buying Z.
- Banners with discounted products in the same category converted better than those with product X at full price.
- Most users who clicked on the banner clicked on a top-10 selling product rather than the product they had seen earlier.
The role of big data in retargeting is not necessarily restricted to these areas. As an example, banner click-through data is used to determine and implement frequency caps for every user segment. This implies that people who show a propensity to not click banners or are no longer considered a valuable prospect by the brand will not be bothered by repeated banners every time they visit a website that offers impressions to the brand. In this case, the brand generates and uses a third big data set i.e. the retargeting campaign data itself.
Retargeting offers the world of digital marketing several opportunities to get its big data strategy right. In order to maximize marketing ROI as well as to make ads less annoying and more contextual for users, brands must learn to adopt big data science as a genuine practice. Brands that run their own retargeting campaigns must hire a data scientist to ensure every dollar spent on an impression is scientifically nuanced. If campaigns are outsourced, they must know how much and what kind of big data science will be applied to their campaigns.
Original article published by Vizury Marketing on Clickz.