In what’s likely to be a first for any publisher globally, TMG is announcing a partnership with ClickDistrict to launch a bespoke bidder to power its Publisher Trading Desk. Here Martin Van Der Meij discusses the new solution, the ad trading strategy at one of Europe’s largest publishers and why we are likely to see more European publishers going down this route.
You have easily been one of the most progressive publishers in Europe around automated ad trading. Do you think the demand and the technology exists to allow a big European publisher like TMG trade inventory in this way?
Not yet. Technology on the demand side is built primarily for the demand side. It always has the best interests of the advertiser, agencies and trade desks, who operate on the demand side of this industry. With a publisher trading desk we are moving in the direction of the demand side but we’re still a publisher. We want to increase the advertising experience for our users. Yes, like every publisher we make money from the advertisers. But our primary business is content for our audiences. It’s about a better user experience.
How do you handle so many websites on your SSP? Have you had to tailor your solution around this?
It’s the biggest issue we had over the last couple of years. What’s the perfect setup? How do you handle floor prices? How do you keep a close relationship with the publisher and our visitors? What kind of data should be made available and for whom? We came to the conclusion we had to split our SSP up into thee new buckets.
You said you have split up your supply into three different Automated Trading Platforms for TMG Can you give some overview on how these ATPs work, and how they have beenefited publishers?
With the current state of the market we “rebranded” them to Automated Trading Platforms (ATPs). Yes it’s still SSP technology, but we’ve redefined our approach. The first ATP we launched was ATP-Social.
For social network sites the URL info that’s passed to the DSPs is close to useless. So we’re in the process of developing a new setup. URL data will not be leading but profile (user) data. The focus is on profile data instead of websites/section etc, we’re focusing on tradedesk and agencies instead of “adnetworks”.
And of course the main component here is the A in ATP. So it’s automated trading only. This way we have full control over revenue streams. From a sales perspective we can actually see how advertisers are spending direct vs automated.
Within a few months we’ll also have our ATP-Classifieds running. Based upon cross-site channels (example: all automotive sections on several websites) and intent data.
Our current SSP will be rebranded to ATP-News before the summer and consist of national, regional and local news websites. Making it possible to sell local ads at scale.
In short advertisers and agencies (trading desks) get three options. You want to buy reach, profile data or intent data? We’ll have platforms for all three and our Publisher Trading Desk can offer cross-ATP combinations. Where can you reach exactly the right audience with the right intent on the right inventory at scale? We will be that publisher in the Netherlands.
You’ve been pretty vocal about the whole seperation of remnant and premium inventory argument in the automated channel. Do you think the majority of publishers are taking the wrong approach on the ad trading piece?
Every market is different in Europe. So maybe it makes sense in some markets to have that brick wall between premium/remnant. Sometimes you just have to agree to disagree. And, yes, I disagree big time. I’d rather talk about automated vs non-automated.
From a user perspective there is no difference between premium and performance and/or remant. Advertising is relevant or it isn’t. It’s either annoying – or not. To make sure the user benefits as well as the advertiser you have to merge these two.
The difference between premium/remnant is based on what? Price? Maybe the difference should be based on relevance for the user? Or ROI for the advertiser? A Homepage Takeover is “premium”. For whom? The user, advertiser or publisher?
From a sales perspective it makes total sense. You don’t want to automate everything. You need those conversations with the advertiser and agency. But then automated trading is not about “not talking” – it’s about how can we make it as easy as possible to connect the right advertiser with the right user.
Do you have any idea how much conversation you need with an advertiser to make sure you achieve just that? Automated trading is about efficiency. Not about silence and ignoring the voice of the advertiser.
TMG was the first publisher in Europe to launch a Publisher Trading Desk. You are now looking at launching your own bidder. What advantage will your ad clients get from the TMG having its own bidder?
The biggest issue everyone using a DSP technology has right now is the internal auction. So even before a bid is on our side (one of the ATP’s) we needed to win an auction (within the DSP). Competing with almost everyone. And of course then there’s the risk of losing our bid on a DSP level while the winning bid on the DSP level is being rejected by ATP (floorprice/block etc). From a technological and cost perspective it makes sense.
But for us as a publisher we needed something else. Much more control of what’s happening with our direct advertiser deals.
As a publisher I want to have an advantage over all the other trade desks. Especially on our own inventory. Our own bidder was the only logical solution.
How does a publisher build a bidder that fits its needs? What’s involved in this process?
Building a bidder is not our business. So we start with a white labeled version of “Platform161”. Marco Kloots (CEO of Clickdistrict) and I go back for years, so yes I was excited when I knew he was working on his own technology. The Dutch market is different then the USA market and we could see how this system was performing over the last 6 months on our ATP’s.
Another benefit of being a publisher is we can compare DSP’s on their outcome. How much revenue do they generate? Whats the win rate? eCPM etc. Not on a advertiser or tradedesk level – but DSP. That’s technology.
And for us it’s just a start. Right now we’re in the middle of the “whats next” question. How can we leverage our data into our DSP technology? What changes do we want? We’re not talking about how to use the technology. How can we alter the technology would be a much better question.
Remember SpecificMedia buying MySpace? Basicly we’re going in the opposite direction.
Will you use two different vendors on the buy-side and the sell-side? Is that natural friction required?
Every buy side technology has the best interests of the buyer. Every sell side technology the publisher keeps is on the side of the publisher. There’s your conflict.
One is about margin – the other about media cost. Our Publisher Trading Desk team is competing with the guys who ARE running the Automated Trading Platforms. Both need to drive revenue up.
I’m still proud of the fact that our PTD team finds it hard to spend all of our budget. That means there is enough competition and it forces us to keep improving and innovating.
And it’s not more expensive to have several systems. That’s just another myth. We’ve done a comprehensive verification of our internal set-up. Cutting costs was impossible. Every change we would have made would have led to less friction, and more revenue for one of our vendor partners – with no overall benefit to us.
Yes we pay multiple fees over one impression. But its about ROI. Not cost. Extra costs are not always a bad thing. Just look at the additional revenue you can get.
Given your access to your own proprietary first party data, could the TMG bidder become a competitive buy-side solution for agencies and advertisers?
Yes. For three simple reasons:
1) We have different technology compared to at least 50% of the buyers in the Netherlands. And more importantly, we’re able to set the roadmap in development ourselves. We’re not dependant on 3rd party road maps anymore. We can build tailor made solutions for advertisers.
2) We have access to data on a page view level while DSP’s only can grab data per impression. And not just page view. We have a huge amount of data within TMG, such as offline, subscription, email, etc, etc.
3) We’re talking to the advertiser and agencies for non-automated and automated budgets. So we have a whole new conversation with the advertiser. How can we help them make it easy to buy standard display? And how can we combine all automated buys with things that need to be handled manually at this moment?
Should all big publishers be looking at building their own bidders? Will this strategy only work for closed markets like the Netherlands?
The DSP auction problem is bigger in the Netherlands then other markets I think. Not because of the DSP/SSP technology. But other markets might have a better diversified market – more DSP’s/SSP’s with a better share of the market. Maybe I’m a year ahead. But a different bidder per publisher trading desk? It would make sense. And it won’t just stop with PTDs. I think advertisers will start building their own bidder.
You will be sitting on “The Rise Of The Dynamic Sales House” at the Ad Trader Conference on April 19 (moderated by ExchangeWire Editor, Ciaran O’Kane). Given that the German Sales Houses are very strong in that market, what advice could you give these publishers looking to build solutions around the automated space?
Make sure you’re ready for automated trading. Don’t adopt a standard strategy, but have your own. You said it in your first question: TMG is known for being progressive. If you want to make any headway in the dynamic marketplace, this is the only approach. As for cannibalisation of revenue? It will happen anyway. Just make sure the new revenue is going to someone within your own company.