If you’re a publisher and considering using a Data Management Platform (DMP), you are most likely thinking about ways to enhance your ad sales proposition. Fortunately, you’re in luck, because DMPs can definitely help in that regard. But data is fluid, so shouldn’t it be applied elsewhere and everywhere?
Let’s take a look at several ways to obtain value from a DMP if ROI is a concern for you, splitting these methods out by quantifiable and less tangible benefits.
The use-cases most publishers focus on are typically those with hard numbered KPIs that can be tracked. Direct correlation between selling an audience and ROI is quite easy to prove.
The prototypical approach to using DMP as a web publisher involves leveraging data collected from your online presence and/or available second- or third-party data to add to your ongoing direct sales package marketed to agencies and brands, leading to increased ad sales.
While this is the typical reason publishers use DMPs, often times these companies forget that the same can be done programmatically in PMP or Programmatic Direct through an SSP rather than an Ad Server.
Regardless of whether these audiences are offered to the market through an ad server or SSP, the general approach is the same:
The same audiences built for ad sales can be exported to Paid Media channels (DSP, Search, Social, etc.) in order to increase the reach of your audiences, also known as Audience Extension. If the available impressions for your top earning audiences are maxed out, but you still have advertisers looking to target your users, a potential opportunity lies in pushing the audience to a DSP to target the same group of users offsite.
Or maybe the advertiser would like to target your audience on THEIR inventory or a whitelisted group of sites they have in their DSP. Again, by exporting your audience beyond the ad server will allow such a use-case.
It’s not just display that these audiences can be applied. They can be exported to any digital (and potentially non-digital) channel to target on search, social, email, SMS, etc. By extending the reach of your audience, you can increase the size of campaigns delivered to increase overall ad spend.
Now that your audience has been pushed to various channels, and you’ve seen the success of advertisers targeting the users without your inventory in tow, perhaps it’s worth looking into selling this data to the broader market.
Whether you choose to sell via direct relationships with buyers or by making the data available to any buyer through a DSP data provider seat is your choice. Alternatively, if you don’t want to sell under your brand, maybe it’s worth a look joining an anonymized data exchange. All of these options are a chance to earn income passively using an asset you are already collecting with your DMP.
We’ve explored options where your data is formed into audiences and provided to benefit buyers through ad serving, but isn’t there any benefit of this data to yourself?
Value from data and from a DMP isn’t necessary quantifiable. When a sale is made to an advertiser, it’s not difficult to see the tangible revenue metrics. Mapping inherent value of data for personal use, however, is not always so easy.
Your DMP of choice should provide the ability to understand audience overlap with the raw first- and third-party data available in the platform. This will provide key insights into how your users behave. If your DMP doesn’t offer such insights, try inquiring whether a custom query can be run (by the DMP or yourself) such that similar results are obtained.
Through the previous publisher use-cases of the direct sales-based ROI type, you will have built out several interest-, intent-, and/or demo-based audiences. These are helpful to understand overlap from, as it will help your ad sales team tell a more complete story to the advertiser differentiating your users and your data from market competitors.
But, to add another layer, try building out audiences based on your site visitors (site, section, or URL) or specific site interactions (watch videos, shared content, filled out forms, logged in, etc.). Running an overlap report on these types of audiences with first- and third-party data will provide an incredible wealth of interest, demo, and potentially intent data against your site consumers that may not have been known or even hinted at previously.
For instance, an audience of users who watched videos at least 3 times in the last 5 days might show that 70% are men, 40% are age 25-34, 35% are interested in sports, and 20% are in-market for automobiles (as well as overlaps with all other data points)! These are powerful insight.
Imagine how the marketing team can use such insights to adjust how they approach potential content consumers, or the sales team working with new advertisers in verticals previously untapped. The limit of the uses of your audiences is only capped by your imagination!
Consumer Insights obtained through the DMP can be applied to adjust your site content in several ways.
From the previous example, you know that 70% of people watching videos at least 3 times in the last 5 days are men, 40% are age 25-34, and 35% are interested in sports. Perhaps the video producing team should make a few videos about football (regulation or American). Knowing the makeup of your audience allows content to be built specifically for them. When content is relevant for individuals, it keeps them engaged and more likely to be loyal consumers.
Customization helps to modify the types of content your editorial team produces, which helps to narrow in on a few broad categories. But how can we ensure that the user is getting the most relevant content to their specific desires?
Let’s say that 1 out of 10 articles were relevant to the users coming to the site. With Customization, we can say 4 out of 10 articles are now relevant. How do we increase that percentage?
Instead of making articles more relevant (i.e. customizing the content) why don’t we reduce the irrelevant ones so they aren’t displayed to users who are not interested and only show them to users who ARE interested?
We know from the Consumer Insights that many consumers are young men interested in sports. What other major audience groups are there that make up consumers’ collective attributes (and those who aren’t regular consumers)? Well, women still make up 30%, which is a large chunk. And maybe some other important interests beyond sports are ‘finance’, ‘fashion’, and ‘politics’. Therefore, the editorial team can build out content around these topics as well. But how do we ensure the right articles are delivered to the right users?
That’s where Content Personalization comes into play. Through a CMS engine, DMP audiences can be fed to ensure only the most relevant articles are getting displayed to the current user.
If a user has a finance audience in DMP, when the user visits the site, a finance article is presented on the home page (through the CMS). If the user is a member of both the finance and sports audiences, does the user have different frequency or recency for these? If they are part of finance 1 time in 30 days and sports 3 times in 7 days, then sports is more relevant. (Or perhaps you as the publisher are weighting one vertical over the other?)
Now instead of 4 out of 10 articles being relevant (from just using customization), it’s 4 out of 8 articles that are relevant. It’s not perfect, but we’ve gone from 10% relevancy (1 out of 10 articles) to 50% (4 out of 8), an increase of 500%!
So now we know how to ensure the right content is shown despite the users’ interest or demographics. But we can’t just keep showing the same article or type of article to the same user.
Content Recommendation can help solve for that. As a user browses the page, to one side or at the bottom of the article a selection of relevant articles should be offered to the reader as additional content to consume. If the user is reading an article about sports, but also has finance and politics in their DMP profile, perhaps the list of articles being recommended is 1 article each of sports, finance, and politics. Integrating the DMP with a content recommendation engine (proprietary or 3rd party like Outbrain) will allow these relevant articles to show based on the DMP audiences.
We weren’t perfect in making every single article provided to the user be relevant through Customization and Personalization, but by offering the consumer a choice as to what content to read, we can help minimize that inefficiency.
The typical use case of a DMP for a publisher is ad sales or an extension of it via audience extension or data monetization, as these are both easy to track in terms of immediate ROI. It’s easy to tell how you stand when you get a check at the end of the month for a campaign delivered or data sold.
Quite often publishers miss out on an even easier opportunity, which is using the DMP to benefit themselves! And ROI is as easy to prove as monitoring the uptick in user dwell time and/or subscribers to your service!
The reason this opportunity is missed may be, unfortunately, due to the internal executive sponsor of the DMP being from Sales or Ad Ops rather than Marketing, Consumer Insights, etc. Why would Sales, for example, directly care about content? Sure, the case can be made that the longer a user stays on the site the more attractive the inventory, but directly it’s not something felt by the sales team the way audiences being applied to Ad Sales is.
My recommendation? Find an internal resource who is looking for opportunities across the board to take ownership of the DMP. It’s important to have all relevant teams on-board with DMP – what it is and how it will help them personally. Otherwise, without any traction you’ll be spinning wheels.
Regardless of who the primary owner is, though, if you’re working with a DMP and find yourself in a rut, take heart knowing that the use-case you’re working on isn’t the only one. Value should be found for every team touching digital. Because data is fluid, it can and should be utilized across every part of your organization for maximum impact.
By Hunter Terry, Director of Global Innovations, Lotame