How Advertisers Get Customer Data and What They Do With It

In our previous article on ad targeting we highlighted the clear marketing trend away from broad ad targeting toward increasingly narrow ad targeting in order to increase return on investment.

Besides the ability to reach users with an ad which is highly relevant to website user (contextual targeting), advertisers are able to reach interested users at any website they’re on with a personalized message, based on users’ interests and browsing habits (behavioral targeting). Marketers adjust their message depending on the audience segment with a striking precision that can only be achieved by collecting anonymized customer data and deep data analysis.

That’s not all, however. Advertising is becoming increasingly automated—and increasingly precise.

Let’s take a closer look at how your data is used, how ad tech companies like Data Management Platforms are involved in its analysis, and how this all contributes to further automation of digital advertising.

The True Power of Customer Data

Customer data collection and management has become an obligatory prerequisite for any marketer who is serious about connecting with his users and reaching desired marketing outcomes.

If advertising is the “lifeblood of internet”, information about customers and their online habits is the “lifeblood” of successful advertising.

To achieve the best possible results, businesses have always been on lookout for solutions that offer a “full view” of their customers. Businesses are eager to know how customers  communicate with their brand and which devices and marketing channels bring the most value in terms of customer acquisition and retention.  Of course, advertisers accumulate certain data about their customers themselves, but this manual work is hardly enough to create a “complete picture” of customer persona.

If advertising is the “lifeblood of internet”, information about customers and their online habits is the “lifeblood” of successful advertising.

This survey shows that 94% of businesses agree that marketing personalization as critical for their success.  A 2017 Holiday Retail Outlook Report by Conversant Media shows that 65% of customers are fed up with irrelevant brand communications.  

Luckily for brands, the ad tech industry has a solution to this problem—Data Management Platforms (DMPs) and related entities like DSPs (Demand-side Platforms) that power ad buying in an increasingly automated way.  

What Are DMPs and How They’re Involved with Data?

List of ad tech companies involved with customer data
List of ad tech companies involved with data. Source:

Above is a comprehensive list of ad tech companies as of May 2017. There is a whole column listing companies involved with data, doing everything from analytics to visualization. There is also a section for DMP companies which includes giants like Oracle, Adobe, Exelate and number of smaller ones. Some of these companies are responsible for the daily dose of ads finding you across the web every day.

Here’s why.

Data management platforms have several key functions, including collecting data from a range of sources then storing and organizing it to provide advertisers with insights regarding their audience.

By design, DMPs are intended to process, normalize and sort massive data sets—a task that can’t be completed by advertisers on their own, due to lack of resources, infrastructure, and sheer enormity of the datasets.  If you’ve heard of “big data”—this is exactly what DMPs are all about, since they not only collect data but create an added value through the generation of actionable insights out of unstructured information.

What’s it all for?

The entire business model centers around the growing need for advertisers to improve ad campaigns efficiency and reduce ad spend—continuously. Demand for DMP services is growing steadily. According to a study conducted by Persistence Market Research, the global data management platforms market is expected to reach $3.7 billion USD by 2024, demonstrating 14.5% compound annual growth rate from current mark of over 1.2 USD billion.

Where Do DMPs Get Customer Data?

First-party data is supplied by advertisers themselves and is known to be the most valuable for marketers, since it gives a ton of information on existing customers, which DMPs leverage to help brands find more customers similar to those who are already loyal/buy their products (lookalike segments).  First-party data includes:

  • Data the advertiser collected from their own site through analytic tools (Mobile/Web)
  • Ad campaign info collected by the advertiser (Search, Display and Email campaigns)
  • Transaction data (Registrations, purchases)
  • Cookie-based data
  • Offline data (the advertisers customer relationship management (CRM) system information)

Once an advertiser places a DMPs tag on their site or mobile app, customer data (unique IDs)  will be collected automatically in realtime and imported to the DMP for further organization and analysis.

Second-party data is basically data from other advertisers that can be shared/purchased and used in a DMP. For example, a company selling boats is likely to purchase user profile data from a company selling fishing equipment because these groups are likely similar customers. All of this can be done through a DMP, like Adobe Marketing Cloud’s Audience Marketplace. According to Adobe, they control all the legal and privacy aspects of such deals between the platform’s clients.

Third-party data is gathered by data collection companies as well as DMPs themselves. It may be used to enrich advertisers’ first-party customer data by adding detail that otherwise would have stayed unknown to advertiser. Third-party data is aggregated into segments. It is normally cookie-based and as such is anonymous. Advertisers may purchase selected segments like “people who are in market for boats” from data collection companies and import them to their DSPs.   

What Happens to Customer Data Inside the DMP?

This is the most complicated part, mainly due to the architecture of DMPs and the multi-faceted usage scenarios that differ for each marketer.

Initially, a DMP is integrated with an advertiser’s website for first-party data collection through placing a DMP’s tags. Other types of first-party data from a CRM system or offline purchases may be imported to the DMP as well. This data may contain PII—personally identifiable customer data, like addresses. However, audience data is anonymized and stored by DMPs in the form of unique cookie IDs or device ID’s. At this point, offline first-party data may already be synced with:

  • Marketers’ first-party online data, which may help to create behavioral segments of existing customers according to marketer’s site usage or of visitors with intent to making a purchase.
  • Third-party online data, like cookie IDs, from the ad networks an advertiser uses. This is possible because DMP tags allow cookie syncing.

DMPs import and process organized data clusters from all parties/sources, both online and offline, in order to consolidate data and compare marketers’ first-party id’s with data points from third-party sources. This process enriches advertiser data with a range of information regarding the audience—from device usage habits to detailed demographiс metrics.

How DMP works with data
A scheme of how DMP works with data

DMPs take it a step further, however.

DMPs can derive new data batches that provides a view of the advertiser’s audience in the form of customer profiles. In order to understand the granularity of customer data available to marketers through DMPs, take a look at an example of customer profile from a presentation by  IgnitionOne DMP .

This sneak peak into marketers’ realm reflects thoroughly recorded and organized data on a customer’s  device usage patterns and online behavioral stats. Even more—they know this possible customer has a car! Talk about an all-seeing eye!  

Example of customer profile based on first-party and third-party data.
Example of customer profile based on first-party and third-party data. Source IgnitionOne DMP

Additionally, advertisers can create audience segments based on common traits between user profiles in order to target segments at the right time and through the marketing channels/devices which have highest probability to deliver business results.

For example, an advertiser may create a segment consisting of women aged 30-55, who open at least 50% of newsletter emails and visit the website at least once a month. They use their mobile phone as primary device and the advertiser wants to market new mobile app to them.

Here’s another example of segments that can be built inside the DMP,  based on how users interact with marketer’s website:

Audience segments based on interaction with marketer's website
Audience segments based on behavior (interaction) with marketer’s website

Marketers may use these segments to set up campaigns to engage each segment through ads, placed on other sites (this method is called “retargeting”).

Audience segments defined by DMP’s are used by advertisers to improve programmatic ad campaign results when they are bidding on ad impressions sold through Demand side platforms and Ad Exchanges. Data exported from DMP’s to demand side platforms improves quality of media buys and conversion rates, consequently reducing ad expenditures.

DMPs are supplyiing DSPs with information required for smarter ad purchases in programmatic media buying.
DMPs are supplying DSPs with information required for smarter ad purchases in programmatic media buying. Source:

The possibility of reaching customers across devices and various marketing channels means the ads are becoming omnipresent —a brand’s message reaches you anywhere and everywhere.

Remember we mentioned how marketers need to continuously adapt and improve their campaign strategies? DMPs have it covered.
Since DMPs provide ongoing analytics of campaign data, marketers are able to single out the best converting profiles and build new segments of audience to target their ads to. This scheme will give you an impression of the role DMP plays in this highly automated process.

What Are the Consequences for You as a User?

You may already have understood some of the possible consequences just from reading about the data collection and usage, but for clarity let’s break them down.

Advertisers don’t just know they can predict with startling accuracy what you’re in the market for. It’s not that advertisers know what you’re doing online; they predict your next step to certain extent and are ready and waiting to arrest your attention. Some of the most sophisticated algorithms are constantly at work, analyzing piles of offline and online data to predict your steps.

Furthermore, using machine learning, these algorithms are getting “smarter” at astounding rates. Computers don’t sleep!  

Advertisers don’t just know what you might buy, they can target you during Every. Step. Of. The. Buying. Journey. n fact, based on your segment in advertiser’s DMP,  you may start seeing ads whose message alters as you move from showing interest to considering certain product through to eventually purchasing it. This is known as sequential messaging. In case you are distracted on your journey to purchase,  advertisers may use Retargeting  to prevent you forgetting about them. All of your interactions with the brand (ad views, clicks) can be stored and analyzed by a DMP’s algorithms. When remarketing kicks in, ads will find you across multiple websites, reminding of the “unfinished  business” you have with a certain advertiser.

This brings us to the next aspect:  

Advertisers don’t just know where you are, they know where you are on each device you use.By implementing cross-device features, DMPs greatly empowered advertisers. This feature allows advertisers to build audience segments across device types and use this info for ad campaigns. The possibility of reaching customers across devices and various marketing channels means the ads are becoming omnipresent —a brand’s message reaches you anywhere and everywhere.

Advertisers don’t just know when you’ve changed your mind, they adapt to your new preferences. You may notice how quickly certain advertisers can adapt to your behavior and roll out a new offer or find you through other website or device. This is because the advertiser has a DMP at his disposal, offering analytics, campaign optimization in real time, and audience insights in one place .Such flexibility is in no small part due to machine learning algorithms that supply a consistent stream of new actionable data for advertisers.

If this all feels a bit invasive. It is.

How Should Customer Data and its Privacy Be Protected?

Regulations like the GDPR in European Union and organizations like Digital Advertising Alliance in US and Internet Advertising Board internationally, set rules on how users’ data should be treated. There are several basic aspects related to privacy that are worth noting:

  • Sensitive first-party information about customers that is uploaded by an advertiser to a DMP needs to be properly anonymized. This is especially true of offline or CRM data, which may contain Personally Identifiable Information (PII).
  • As a user, you should be notified of DMP tags and cookies on any website and be offered a way to opt out of data collection.
  • Some DMPs use cloud solutions for storing advertisers’ data, but it is preferred that advertisers use on-premises storage. There are several DMPs that work on a Software as a Service basis, allowing advertisers to deploy an in-house DMP.

What to Do If You Feel Uncomfortable with Too Much Targeting?

As mentioned above, opting out is a good solution. Using a comprehensive list of DMPs and ad networks compiled by technology compliance company TrustArc, you can do a massive opt out of data collection.

In addition to the opt-out, here are some other helpful tips from the first article in the series: