In my last article, I covered the foundations of programmatic planning principles. Today I will focus on the biggest issue in the Chinese programmatic ecosystem: conflict of interest.
There are many disciplines within the programmatic ecosystem, some focus on serving the advertisers, such as through demand side platforms (DSP), while others primarily serve the sell side, such as exchanges / supply-side platforms (SSP).
In the western programmatic ecosystem, most vendors have an exclusive focus that rarely crosses into other disciplines. But in China, it is the exact opposite, as most of the programmatic players tend to offer one-stop shop solutions that spread across the entire programmatic stack.
For example, this graph shows China current programmatic landscape. To illustrate the point of conflict of interest, you can see that quite a few companies show up in multiple categories.
While it may seem convenient that a single vendor can take care of a brand’s programmatic needs, advertisers will in turn lose a lot of transparency due to the massive conflicts of interest between a vendor serving both the buy-side and the sell-side.
Managing Conflict of Interest
In order to minimize transparency risks, I’d like to layout some of the most common conflict of interest scenarios in the Chinese market:
Scenario 1: The Buying Platform Should Not Own the Ad Network Inventory
This happens because most DSPs in China started out as ad networks, which works on a mark-up basis.
When the programmatic market started to grow, the traditional ad networks didn’t want to instantly make the transition to exchange inventory. Hence, the term ‘DSPAN’ was coined, meaning a hybrid model of an ad network plus an exchange.
This model fulfills the needs of a more traditionally minded advertiser that may still want to pick from a fixed ad network publisher list.
On the other hand, it can also cater to the more performance-oriented advertisers that like to buy inventory through the exchange.
However, when the two inventory sources are sold together to the advertiser, then it is in the DSP’s best interest to sell their own ad network inventory first prior to buying on the exchange. In this case, the buying platform is inherently biased in their inventory pool selection.
Scenario 2: Separate the Buying Platform From the Data Provider
Data is the core of programmatic, providing the buying platform with audience tags.
Since there is a lack of third-party data in China, most of the DSPs also have a second-party data management platform offering. But if the DSPs make a mark-up, then their second-party data will justify buying ads on the exchanges, which offer the highest margin – not what’s right for the client.
Hence separating the data and the buying platform will give DSPs less justification to buy junk inventory.
Aside from the issue of mark-ups, there is also data security and privacy at stake.
Some of the biggest brands have accumulated large amounts of CRM or offline sales data that they also hope to use as the base of programmatic selection. This dataset is way more sensitive than data collected on digital assets, because it represents the brand’s most loyal consumers.
Hence, it’s even more important that these datasets don’t get integrated into a DSP’s second-party data, where they can use it for competing brands in the same category.
Scenario 3: The Buying Platform Cannot be a Media Owner Then Act as a Broker For Other Publishers
In the publisher-dominated ecosystem in China, some dominant market publishers can offer complete programmatic solutions. The first part of these solutions will be a publisher-owned DSP product, which connects to the publisher’s private exchange, enabling advertisers to buy both guaranteed premium inventory as well as biddable remnant inventory.
The second, and the most enticing part of the offering, is when buying inventory this way, the publisher provides access to their DMP based on a combination of cookie-level browsing behavior data, as well as consumer provided registrations data.
This is a valuable offering in the market due to a general lacking of third-party datasets.
The third piece of the solution entails a private exchange that also allows advertisers to source inventory from other publishers. This complete solution may again be very convenient for advertisers, but the stack is wholly owned by a single publisher.
In this case, not only does the publisher lock the advertiser down with data ownership, but they also act as a broker to buy inventory from other publishers.
This solution essentially gives a publisher entire control over an advertiser’s digital media budget.
It is of the utmost importance that conflict of interest is understood by clients in China.
Next month I move into tech stack evaluation and recommendations for ways to evaluate the right buying platform.