IAB Europe’s Virtual Programmatic Day took place on 28th April and with 1,500+ attendees it is one of the largest virtual events in our industry. Featuring speakers from IAB Europe member companies CNN, Bloomberg Media, BBC Global News, Double Verify, OMD, MediaCom, Google and OpenX, the event was split into panel sessions to address and debate the industries hottest topics from post-cookie to programmatic in-housing.
The State of Supply Path Optimisation Panel
The third panel of the event addressed the technicalities of supply path optimisation and whether it has delivered on its promise to streamline buying via programmatic platforms. The panel also looked at this from a publisher perspective addressing demand path optimisation (DPO).
Daniel Knapp, Chief Economist, IAB Europe moderated the panel and was joined by:
Event Recording
Watch the event recording here.
Audience Poll
We ran an audience poll during the panel to understand what has driven the demand for SPO and revealed the key driver, according to the attendees, is the demand for more transparency. However, most think it is driven by a number of reasons including change in auction dynamics, demand for transparency and pressure from regulators with half of the audience voting ‘all of the above’.
Audience Q&A
The panel received more audience questions than there was time for. The panellists have provided their viewpoints on some of those questions here:
Q. Advertising is a publisher's core-business. They should have been leading transparency since the very beginning. How do you think all these transparency initiatives will impact on publishers results? Why do you think it took so long to get started?
Tom Fryett: You could say advertising is an agency’s core business too. I think buyers and sellers were always trying to find one another and give each other transparency to foster trust - but also ultimately to do more business together. The tools improve over time. We derived a huge amount of insight from IAB initiatives like Ads.txt which has been enhanced by Sellers.json and the supply chain object. We have the resource and tech on our side to analyse this data at scale and use it to make investment decisions. I think truly premium publishers will benefit and if they have a scarce and desirable audience to advertisers, they will be rewarded by this work on both sides to have transparency.
Simon Baker: Advertising was a publishers core business but subscription and paywall revenue have overtaken ad revenue in multiple examples and that's excluding live events.
Regarding transparency, I don't think either the buy side or sell side set-out intentionally to create a murky world - it's something that's naturally happened with multiple players in the programmatic ecosystem. If the question is should publishers have done more to protect their inventory from the outset then the answer is yes. The awkward truth is bad decisions are sometimes made when focusing on short term revenue pressures.
Daniel Knapp: Transparency by itself does not have value. The ability to look into a system needs to be paired with the ability to make sense of it, and then to take appropriate action. This is hard to do, and most publishers are not set up for this. Algorithmically traded markets, whether in finance or advertising, remove institutional knowledge about how the market operates from the ultimate buyers and sellers, and even from intermediaries.
Q. Can you please elaborate on inconsistent auction dynamics?
Shane Shevlin: This refers to the type of auction that an SSP runs for OX traffic. It can be first, price, second price or it can be something in between. Jounce Media’s “RTB Supply path Benchmarking Report” from last year (2019) reported that the average publisher transacts open auction demand through 17.5 distinct supply paths. Each of these paths will result in different economics, with some paths offering much less value along the way. This both affects the planner/buyer putting their budgets on the line with extra fees for longer paths and the platforms who incur additional cost managing the extra bid listening costs, managing slower response times etc.
Q. Will publishers need SSPs in the future? Isn’t the market working hard to skip at least one -SP, be it the D or the S?
Tom Fryett: We love the “Death of the X” metaphor in this industry. DSPs and SSPs will evolve to meet new challenges. They won’t go away completely but they won’t necessarily look the same in the future. It will be driven by the economics of demand and supply - TV content is very expensive to produce, has long lead times and you tend to have fewer, premium ad placements around it - therefore DSPs and SSPs will evolve to make the process of buying that type of supply more efficient, rather than shoving it into a Display way of working.
Krzysztof Lis: I don’t think that this will be the case. If a DSP wanted to integrate closer with the publisher, it would eventually mean that the DSP created it’s own SSP, right? The only reason to implement this DSP/SSP hybrid would be to get the new, additional, unique demant that comes from this DSP. But from that hybrid’s perspective it would be beneficial to open their inventory to other DSPs as well -- so we’d end up pretty much in the same place, as we are now.
Lisa Kalyuzhny: We covered this question during the session - this was a hot twitter topic 5 years ago - and seems that people are keen to bring it back up. For us to live in a transparent environment with an honest holistic auction - we need both sides of the equation and for those sides to work together not be a single company. Otherwise your auction is no longer a “fair” auction - you are favoring your own demand or supply - and that hurts both the advertiser and the publisher.
Daniel Knapp: Publishers will always need specialist partners that can manage and tap into demand in a programmatic world. The resources and scale required go beyond what a single publisher can provide in-house. But a few years from now, this may not (only) be a classic SSP function as we know it. The beauty of the programmatic marketplace is the democratisation of supply and demand. But the economics may not add up in the long-term. The Walled Gardens partly excel because they compress supply and demand technology into one offering. Outside these gardens, there is a drive to maximise working media and also to differentiate as a tech provider - for instance through exclusive relationships. We already see direct relationships between DSPs and publishers, and also between agencies and SSPs. There is a risk of the programmatic pipes becoming less open and more proprietary. The benefit is a potential world where technology becomes a more utilitarian infrastructure than it is today, akin to a software service layer like in CRM. But for the sake of a fair market, this can only work if monitoring and audience systems are in play - a kind of ‘net neutrality’ or ratings agency like Standard and Poors for programmatic pipes. We will see a lot of changes in the Lumascape over the next few years. Yet far from meaning the demise of the SSP, it means that everyone in the value chain will need to innovate to create a more advantageous market position.
Q. Is demand path optimisation (DPO) not just regressing to the old waterfall world where the publisher makes the decision on which SSP they want to prioritise and creating a more fragmented market for buyers?
Krzysztof Lis: Hopefully not, as the waterfall did not have many advantages over header bidding and other currently popular implementation methods. Also, one can easily add different priorities to specific demand (using deals) in header bidding. I think that would be the preferred method of doing it.
Simon Baker: No, I don’t think DPO is a regression. The old waterfall world basically consisted of plugging in every SSP, ad network, reseller and anything in-between and hoping for the best. DPO is a data led approach so we can understand exactly the value each partner is bringing to the table especially from a unique demand perspective.
Daniel Knapp: I see DPO as a key innovation for publishers to take control over their inventory. It is not a regression to the waterfall because it does not imply putting different SSPs into a fixed, hierarchical order. Different SSPs have different strengths. The type of demand they bring the consistency and diversity of demand all vary by time of day, number of relationships, etc. Real DPO is using data science to be the conductor of an orchestra of SSPs: allowing each to play out their strength. In addition, DPO creates data and insight. This business intelligence is critical for more informed and constructive discussions with publishers and SSPs on how to best evolve their partnerships. It also helps publishers understand their buyers better, arm their sales teams and create better offerings. Of course there is a difference between ‘buzzword DPO’ and ‘deep DPO’ mentioned here. But done smartly, it is a far cry from waterfalls.
Q. How do publishers see bid shading by SSPs?
Simon Baker: Bid shading is something that's been happening on the buy side under various different names for years. You could argue there are transparency issues but it’s a tactic that most buyers are using either publicly or privately to deliver the most effective rates for their clients. I would expect most publishers are seeing higher clearing rates than second price auctions so this would not fall into their top five programmatic issues to resolve currently.
Q. Would transparency of QPS cost help publishers and advertisers understand the impact inefficiency has on the ecosystem? If so, would they be willing to pay for this bandwidth?
Tom Fryett: QPS cost is a technical challenge that has real-world impact on the shared interests of user, advertiser and publisher. If you delve into SPO you learn a lot more about the infrastructure of the ecosystem which enhances the way you plan and buy. It's a good thing for stakeholders to understand the different challenges different parts of the chain might be experiencing to inform working with one another, even if it does mean learning another acronym.
Q. Do you think that in the future there will be just one technology, which will have the functionally of SSP & DSP, a kind of hybrid of DSP & SSP? Which will essentially limit 350 Ad Exchanges and X-number of DSPs?
Tom Fryett: Further consolidation in the space is a likely outcome of the wider economic situation in the world today, but one technology underpinning emerging over time is unlikely as people will continue to invest in innovation in a market the size of ad tech.
Lisa Kalyuzhny: I agree with Tom - further consolidation will continue - it started with companies closing before GDPR and we’ve continued to see ad tech challenged by regulations and policies. The one technology removes the fair auction and long term becomes detrimental to both advertisers and publishers. It prevents the advertiser from holistically identifying the best impression and the publisher from seeing all possible impressions or interests in its inventory.
Daniel Knapp: I concur with Tom and Lisa. I tried to address my thoughts on this in the comment on if publishers need SSPs above.
Q. With a lot of the new SPO initiatives in place (e.g. supply chain object and sellers.json), how does the panel feel about blockchain companies looking to become another part of the supply chain? Will they go away?
Tom Fryett: Decentralised solutions and smart contracts seem like potentially clever solutions to some of the problems discussed by all sides today. The format doesn’t need to be blockchain necessarily and the simplest solutions are often best when it comes to fostering transparency. Just look at Ads.txt - a humble text file has helped buyers find inventory from authentic sellers - it wasn’t perfect but it was a huge leap forward and has a very low barrier for entry for all involved to use it. Sellers.json and the supply chain object enhances OMG’s ability to verify authenticated partners.
Lisa Kalyuzhny: Blockchain companies were a huge trend in the US about 18 months ago - and then the holding companies realised that they didn't need another hand in the ad tech cookie jar. They had the tools necessary to do their own audits - this opened up conversations with SSPs / DSPs and the IAB. Until blockchain can support the QPS necessary for our industry, it's a buzzword without much behind it.
Daniel Knapp: Blockchain lives off the promise of transparency. But there is a problem. A leger creates the potential for insight, but not actual insight. What is recorded still needs to be analysed. If that is really granular, as it needs to be, Blockchain can’t help, it is advanced and costly data science. Blockchain approaches also suffer from an application gap: there is no path to action. In addition, recording transactions in a ledger itself is insufficient data to understand real market dynamics. This requires experimentation: throwing stones in the pond to test reactions. The business case of blockchain approaches for programmatic was unclear as Lisa mentioned: another media model to live off the chain is counterproductive. Yet the thinking behind what became manifest as blockchain approaches is hugely valuable: how to create transparency and a fair ecosystem. Some blockchain companies have pivoted to pursue these core questions with other technologies and business models. This will create more and real value in the long-term.
Q. @Simon Baker: as a publisher, what are you doing to skip this programmatic ecosystem as a whole? Directly talking to brands (or their agencies) and creating great deals based on mutual possibilities and data.
Simon Baker: Firstly I want to clarify we have no intention of skipping the programmatic ecosystem. We are here to facilitate transacting with our clients and agencies from direct IO to programmatic channels - it’s totally their choice. One noticeable trend is most of our direct business is content led where standard display is migrating to an automated world due to the efficiencies involved. Data is not the dividing point between direct and programmatic, a number of our new data led digital products are built with a programmatic first mindset.
