How to Use a DMP for Content Personalization

The web is a wild and woolly jungle of information, channels, and devices.

Nobody knows this more than advertisers and marketers desperate to find ways to personally connect with customers online, which is why they are becoming keener on leveraging technology for content personalization to cut through the digital maze.

The beauty of content personalization platform is that it can allow businesses and brands to boost revenue without spending a fortune on ad campaigns, simply by harnessing the power of technology to become more relevant to their site’s visitors and potential customers.

Although the possibilities are endless, several areas lend themselves more easily to personalization:

• Product recommendations (on e-commerce sites)
• Content (e.g. blog posts, videos, etc.)
• Cross-promotion (to upsell or promote brands owned by the same company — for instance, Procter & Gamble)

The tool behind these multiple possibilities is a data-management platform (DMP). How does it work?

Let’s take a look.

Content Personalization: A Sample Scenario

In order to illustrate exactly how the DMP functions to provide content personalization capabilities, we can examine an example situation — Eatsy, a fictitious website featuring local restaurants and food-delivery services in a large metropolitan area.

First, we will outline what Eatsy wants to do with its website, then we will examine how a DMP can make this possible.

Eatsy Gets Personal

When Eatsy was started, its creators wanted to offer site visitors an easy place to browse local restaurants and food-delivery services, make reservations, and place orders for home delivery.

Their monetization strategy included:

• a monthly fee paid by a food establishment to be listed on the website
• a small percentage of each order placed
• ad space to allow restaurants to promote themselves on the site

With time, the Eatsy owners realized they weren’t generating very much revenue because:

• a) The order amounts were generally small
• b) Not many site visitors were clicking through on the display ads restaurants were serving on the Eatsy home page

To deal with these problems, Eatsy wanted to do a couple of things:

1. Personalize the ads being shown on the homepage and front page of their mobile app to increase conversion rates
2. Offer a way to serve promotions on each restaurant’s subpage (for example, “upgrades” to larger orders or with add-on items like sides items or drinks) to increase the value of the orders

To accomplish this, a few things need to happen:

• Data must be collected about a site visitor’s past purchase history, location, preferences, device and other factors.
• The data must be merged to create an overall picture of each visitor, or at the very least, it must be merged and segmented in order to divide the group of overall visitors into narrower groups.
• The Eatsy site needs a way to recognize when a specific visitor lands on the site, or at least a member of a certain audience.
• There must be a way to load dynamic, personalized content based on the identifiers assigned to each visitor.

How Eatsy Can use a DMP

So how can a data-management platform help Eatsy accomplish its goal of boosting conversion rates and raising order values?

First, it has to collect the data necessary to make content personalization possible.

Collecting Eatsy’s Online Data

In order to make full use of a data-management platform, there must be plenty of data to work with, and it must be the right kind of data. This is where data collection comes in.

First, there is the question of online data:

• Track a wide range of web interactions using a tag-management system (TMS). These tags will be placed on the Eatsy site itself to track when a visitor arrives, which restaurant pages are opened, which menu items are viewed and ordered, and the value of the order.
• The TMS should also place tags on each restaurant’s separate website. The cookies that are served by the tags in both places can later be matched and synced to get an even clearer picture of visitors’ behavior and preferences.
• Because Eatsy also has a mobile app, its DMP will need to collect data through an SDK. Again, data about visitor’s behavior, orders, and location will all be collected.

Merging and Segmenting Eatsy’s Data

Once the DMP has been populated with a significant amount of data both from desktop and mobile visitors, it can start going to work.

By merging the different data sets and then segmenting them based on various factors, the DMP is laying the groundwork for content personalization success.

For example, based on location data, the DMP can break down visitors into groups, which can later be matched to the restaurants nearest to them. Within those location segments, the DMP can further separate them by value of their previous orders.

When first-party (from the Eatsy website) and second-party (from the various restaurant websites) cookies are synced, the DMP will be able to tell if a visitor to the Eatsy site or mobile app has previously visited a certain restaurant’s website.

As more and more data is gathered and merged into the “live” (continuously updated) segments within the DMP, the audiences will become more and more refined. If DMP includes device fingerprinting capabilities, as well as cross-device tracking, the DMP can present an even more accurate picture of individual visitors, above and beyond general audience segments.

Serving Ads and Promotions

Now having used the DMP to create rich audience segments and build as individual a picture of visitors as possible, Eatsy can activate the data the DMP has processed.

Let’s consider a couple of examples:

1. A visitor arrives on the Eatsy mobile app.

The information gathered by the DMP shows that this visitor has made previous orders through the Eatsy mobile app, mostly for a pizza and almost exclusively from a fixed location.

This time however, the IP address shows that the visitor is accessing the app from a different location. Based on this information, it might be reasonable to believe that the visitor is away from home and might possibly be interested in pizza or related cuisine, but not for order-in.

Using the DMP’s data in conjunction with a content-personalization engine, the Eatsy app could show the visitor a list of local restaurants and/or promotions from establishments serving Italian cuisine, including pizza.

Diving into the details of the data gathered and processed, as outlined above, the DMP could segment the visitor into a category of customer likely to be interested in a “dine-in” restaurant (as opposed to ordering for home delivery) since he or she is using the mobile app and in a different location than usual.

Assuming the data leads to the correct assumption and the visitor chooses one of the offers presented, the personalized content served with the DMP’s help has had a significant effect on revenue: the customer easily chooses a restaurant that suits his or her taste, and because his choice is a high-end restaurant, Eatsy benefits from a higher rate on the personalized ad it displayed.

2. A potential customer arrives on the Eatsy website.

This person has often ordered Chinese take-away food through a certain restaurant’s website and usually orders the same dish, which means his or her average order value is roughly the same.

This time, when the customer lands on the Eatsy home page, the DMP recognizes the visitor (thanks to synced cookies between the restaurant’s webpage and the Eatsy webpage). Now it can serve a personalized offer — his or her regular dish…and at a special price!

When the visitor clicks through the promotion, but before placing an order, they are served another personalized offer: a side dish/dessert/large soft drink. This add-on item is not very expensive, but coupled with the discounted dish offered before, would raise the order value to above average for this visitor.

Thus, the ability to segment the visitor, identify him or her, and link that to previous behavior all makes it possible to serve personalized content, and reap the benefits at the same time.