Jenny Chan 陳詠欣
Aug 24, 2016

Conflicts of interest plague China's programmatic market

While every market has its own set of challenges related to programmatic trading, the biggest issue in China is the protracted conflict of interest between local ad tech companies.

Last Update: 12 May 2016
Last Update: 12 May 2016

Editor's note: This piece is intended as the first in a series addressing the issues raised. As it only presents one take on the topic, we invite ad-tech players or others with a stake in the game to respond to this article and the claims made in it. Please write to jenny.chan@haymarket.asia.

The lumascape above by RTBChina shows China's current programmatic landscape, as of May 2016. It may come as no surprise that quite a few companies show up in multiple categories. But issues arise when you play in more than one playground.

Charlie Wang, chief operating officer of ReachMax, was asked by Campaign to compare the western ecosystem with China's, and said the key difference lies in the dominance of partnerships (West) versus the one-stop shop model (China).

Western vendors rarely cross into other programmatic disciplines, but in China, the exact opposite is true.

"This is the root of all conflicts of interests in China, as domestic ad tech companies don’t have a pure, exclusive focus," he added.

That doesn’t mean that western programmatic ecosystems are free from flaws. They also suffer from inherent bias, mainly due to Google and Facebook providing 'walled-garden' services of their own. But are there aspects China can learn from or adopt?

For example, Sizmek provides programmatic direct buying (PDB) services but does so with one distinction that Wang points out as "very important"—the company does not bundle target-audience verification data alongside it.

Instead, it partners with Nielsen's digital ad ratings (DAR) service to provide that data. This partnership model ensures that it is "only sitting on one side of the fence" with its performance measured by another third party, he added.

According to Del Levin, VP of marketing effectiveness at Nielsen China, it is for this very reason Nielsen does not have a DSP (demand side platform) service. With its DAR being used as the industry’s measurement currency, it would essentially own an unfair advantage over all other DSPs.

“Independent measurement is fundamental if we want to move the industry forward,” said Levin. “Only by applying independent measurement can we determine what is working accurately and what is not...and ultimately enable us to perfect the newer programmatic technologies.”


Three common conflicts of interest in China

1. Scenario: DSP + Ad Network = “The real estate agent that owns property”

Imagine going to a real estate agent to look for a new house, and they repeatedly try to sell you a specific property, only to find out later that the agent is actually the owner of that house. This is what the DSPANs (a hybrid of demand-side platforms and ad networks) in China are purported to be doing. It is well known that most local DSPs in China evolved out of traditional ad networks, but when they made the transition to DSPs, their core business models did not change.

So, given that most DSPANs also own media inventory, they may give preferential treatment to their own inventory as opposed to inventory on other exchanges. As a result, advertisers that invest in DSPANs appear to lose neutrality because they will never be sure if the DSPANs are making the best decisions on their behalf.

Examples: HDT, Adsame

Solution: The buying platform or DSP should not own ad network inventory. Reduce a vendor’s service scope to just a single area of the entire ad-tech stack.

2. Scenario: PDB + RTB (real-time bidding) = “The car repair shop with an agenda”

Let’s assume another scenario: We go to a car repair shop for a routine car inspection, and the shop tries to sell you a whole slew of bells and whistles as "additional services”. Most car repair shops don’t make much money from just an inspection, but if they can convince you to change your brakes, engine, or air filter, they would make a lot more, pointed out Reachmax's Wang.

This scenario is comparable to the one above. For most local DSPs, they tend not to choose PDB over RTB inventory as all the pricing negotiations for PDB are done directly between the agencies/advertisers and publishers. Hence, they cannot make a mark-up in the middle like what they do with RTB inventory.

So, once DSPs in China get their clients to do PDB deals with them, they may attempt to add in RTB inventory as a “value-added service”, but it is understood to be a euphemism for media arbitrage.

Examples: iPinYou [Editor's update: iPinYou's CEO has responded in this article], YoYi

Solution: The buying platform cannot be acting as a broker for other publishers. Introduce independent ad verification and viewability controls different from the buying platform.

3. Scenario: PDB + Tracking = “The playferee” (both player and referee)

For this scenario, let’s imagine an Olympics basketball game where a particular player is also assuming the role of a referee. Do you think that the player’s team will lose? It is a rhetorical question. This type of conflict of interest is the biggest of its kind within China’s ecosystem, because of third-party tracking companies that also supply programmatic direct buying.

These PDB services are sometimes packaged as 'ad serving with pass-back' services, which can be potentially confusing to advertisers, who may think they are not programmatic features. But if third-party tracking companies sell PDB in the same way DSPs do, then they are essentially validating themselves.

Even more worrisome is how these companies often promote on-target percentages of their services to advertisers as a guarantee. Can these 'good' results pass media audits if the same third-party company that delivers media-effectiveness metrics also buys media?

Examples: AdMaster, Gridsum

Solution: Separate your buying platform from your data provider. Be sure not to bundle target-audience verification data alongside media buying.
 


Facing insoluble dilemmas?

The complexity of China’s programmatic environment, plainly put by Rose Du, head of digital at OMD Beijing, stems from "each party wanting to step into each other’s space to make a profit".

Agencies are not immune from this charge either, though the basis for decision-making and supplier-selection is touted to be based on the justified pursuit of meeting client demands and needs.

“As an agency, our role is to generate the maximum outcome for our clients within the given budget," said Fabian Kietzmann, chief digital officer at MediaCom China. "For this to happen, we have to be open towards the whole market. We look at individual costs and at total costs (inventory, data and technology) as well as at the final goal contribution from these costs.”

Kietzmann claimed to "always clearly identify to clients wherever and whenever there is a potential conflict of interest", but he added that in the end, it is not so much about eliminating conflicts of interest per se than "being aware of the market interdependencies”, as it is natural that every publisher is using a yield management system to maximise revenue.

Considering workaround solutions

To prevent more dishonest players creeping into China's fast-developing programmatic market, Johnathan Beh, general manager of Performics China, advises advertisers to introduce independent ad verification and viewability controls, into their programmatic processes.

He pointed to providers such as Integral Ad Science, RTBAsia or AdBug as examples. That said, it is important to keep in mind that fierce competition in the China ad-tech space can force business interests to prevail over an agnostic approach in picking vendors.

So what can an advertiser in order to set up an architecture that allows for truly independent checks and balances?

ReachMax's Wang advocates following one central principle when it comes to programmatic planning—reduce a vendor’s service scope to just a single area of the entire ad tech stack.

If clients start drawing the lines themselves, the reasoning is that the players will follow and when clear positions within the ecosystem are defined then only then can programmatic trading be more transparent and clients more reassured said OMD's Du.

"It will also make the Chinese market be more in line with international standards,” she added.

Vineet Arora, managing director of Havas Media China noted that with ad tech, judgment has significance, especially in China.

"As the paradigms have shifted, it is more important than ever to choose the right things to measure," he said.

Arora added that the real opportunity for agencies and vendors lies in delivering real value to clients rather than looking at other slices of the pie.


Followup articles:

Fugetech weighed in to argue for the meta-DSP model as well as reporting deliverables of more dimensions—a response to iPinYou's statement on why being involved in both PDB (programmatic direct buying) and RTB (real-time bidding) is not a "conflict". Also, Hdtmedia reinterpreted its DSPAN (DSP and ad network) dual role as a "stock broker" recommending media buys with upside potential, while Sociomantic advocated for a 'right to audit' clause in programmatic contracts.

Related Articles

Just Published

21 hours ago

Fortnite maker launches an epic games battle ...

Fortnite maker Epic Games turns Apple's 1984 ad against it as it initiates legal action against the "oppressive" power of both Apple and Google.

21 hours ago

Audit assesses social-media platforms' progress ...

From policy enforcement to misinformation, a study conducted by IPG Mediabrands agency Reprise highlights the progress (or lack thereof) social-media platforms are making on brand safety.

22 hours ago

Punk wellness: Chinese healthcare marketing targets ...

Young Chinese are becoming the main target for healthcare products and brands need fun ways to communicate.

22 hours ago

Haier presents, the fridge you can split when you ...

A TVC for Haier Leader by Serviceplan China presents a new product with a very specific use case.