After defining programmatic marketplaces, let us deal with the renowned and less famous participants of the market. In the following minutes, if you are persistent, you will have the opportunity to learn about the differences between the programmatic model and the traditional model, the terms DSP and SSP, and you will also get an insight into the complicated system of the programmatic ecosystem.
Participants: traditional model
Before introducing the specific participants of the programmatic model, it is necessary to take a look at the participants of the traditional model. The traditional online advertising market has three participants in most cases.
Similarly to all forms of advertising, there has to be a customer, i.e. advertiser, who commissions a media agency (if there is one) to design, manage and optimise its future campaigns. The third participant is the publisher itself, who publishes the advertisement, in this case, the banner, physically (for online advertisements, virtually). Of course, in addition to the participants of the traditional model, the use of different technical platforms, e.g. ad server and data provider systems, is required.
The ad server system is responsible for where, when, how frequently and, most importantly, to whom, a given advertisement should be displayed. As the advertiser and the publisher have conflicting interests, two types of ad servers exist: one on the publisher’s side and one on the advertiser’s side.
The publisher’s ad server has to display the online advertisements in accordance with the policies of publishers. This task seems trivial, but it is extremely important, as, of course, publishers want to regulate what kind of advertisements they accept and publish on their own inventory. Furthermore, they prioritise the individual advertisements. Those which generate more profit have priority.
On the other hand, the advertiser’s ad server has a different role, as it seeks to roll out the display volume expected by the advertiser. In other words, the system has to recognise the rules made by publishers and the users who could be relevant in terms of the advertisement, while maximally satisfying the advertiser’s expectations.
Participants: programmatic model
In the case of the programmatic model, new participants, who are also specialised in the supply or the demand side, are added to the model described above. The demand side is served by the DSP (Demand Side Platform), while the supply side is served by the SSP (Supply Side Platform).
DSP is a technological platform that allows advertisers to buy media appearances in a centralised and automated way, through all available Ad Exchanges, SSPs and ad networks. Based on the above thesis, systems should not be imagined as a 1:1 model, as several Ad Exchanges and Ad Networks can join one DSP on the advertiser’s side. They use various SSPs on the publishers’ sites, thus maximising access and diversification.
If we forget the world of programmatic for a while, it is easy to understand why it is worth running a campaign in several different marketplaces instead of a given site (or a given marketplace). In order to understand this, let us return to everyday life:
The simplest example comes from our shopping habits. Let us assume that a given customer goes shopping to the same supermarket every day, because it is the closest to their apartment. Although it is not a large store, our subject can purchase all essential household goods there. Often, this solution is more comfortable, but it might not be the best and most cost-efficient one. If our subject visited all the shops in the neighbourhood, they would probably be able to buy individual items more cheaply, though one by one.
In the world of programmatic, the customer is the DSP, which is not tied to any place or previous preferences. They can always buy the most cost-efficient advertising space (i.e. the user behind the banner).
If an advertiser purchases media only directly (in the traditional manner), involving only the well-known sites, they miss valuable appearances outside the preferred inventory and a lot of useful users. In many cases, advertisers who are sceptical about programmatic, also run their programmatic campaigns in this manner: they tend to use only the inventory of one specific Ad Exchange (e.g. Google Display Network). In contrast, the openRTB model diversifies the advertiser’s investment, i.e. the advertising budget, thus reducing risk and increasing the scale of the available inventory.
And why does this reduce risk? The answer is simple: if the advertising budget is diversified, an unpredictable negative event will not affect the whole portfolio, but only certain elements of it.
The functions of DSP
The Demand Side Platform has to bid for specific appearances delivered by Supply Side Platforms. The winning bid always has to be displayed in the given advertising zone. In the world of programmatic, DSPs have the added value necessary to decide which advertisement has to be shown when and to whom, while observing and meeting the advertiser’s expectations to the fullest, or, in other words, maximising the advertiser’s ROAS (Return on Ad Spend).
Actually, it is a very complex real-time decision-making system, which has to decide within fractions of a second whether an advertisement is relevant for a given user at the price that the advertiser still accepts. Of course, the system performs this without overburdening the ad server serving the website, i.e. without slowing down the loading of the website. In order to ensure this, the OpenRTB standard was created. In accordance with the OpenRTB standard, a given ad has to be served within max. 100 milliseconds.
In contrast, the SSP on the supplier side is a technological platform that can manage media sales and the control over ad networks for publishers in an automated way. The main function of SSPs is to maximise the profit arising from the advertising inventory, i.e. to receive the highest possible advertising revenue. The fact that publishers could not sell their advertising zones, which remained unsold through direct channels, largely contributed to the spread of SSPs. The unsold zones used to feature mainly self-promo advertisements or empty advertising spaces, which did not generate any revenue for publishers. Due to SSPs, they could sell these previously worthless advertising zones, though at a lower price than the list price of direct sales, in an automated way, giving way to the technological development of programmatic.
One of the main advantages of programmatic technology is that it makes the life of media buyers easier, as the identification of target groups and the appearance spaces based on them requires considerably less time than in the case of the traditional model. That is why this technology is called a paradigm shift, as the campaigns are not optimised based on a plan commonly used in the marketing profession. The main goal of a campaign for example is not to reach as many women aged 18-49, living in the capital, who like going to markets, as possible, but to maximise the number of interactions on the website.
Sometimes, usually in the case of awareness and brand campaigns, it is important to be able to identify the sociodemographic factors on the basis of which the campaign can be targeted and optimised.
This is where pre-defined, first-, or third part data segments, available within DSPs and provided by different DMPs, come into the picture. But before going deeper into the special functions of DMPs, let us recall John Wanamaker’s famous remark: “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” In a broader sense, it is a data problem, as we do not have enough information about which elements of the media-mix performed better or worse. One thing is sure: the more data are available during an advertising campaign, the more accurately we can target the campaign.
That’s why we have to speak a few words about DMP. What is it? Why is it extremely useful?
DMP has two roles: to store and sell the advertisers’ data. Consequently, by using DMPs, advertisers can control the data in all online campaigns, getting an overall picture, with which more efficient campaigns can be run. So DMP is a centralised data management platform that collects, organises and centralises data. Our current topic focuses on sales, therefore we will not mention data use within one’s own portfolio.
Based on the above, DMPs provide an option for advertisers that facilitates the targeting of advertisements by using the data which were collected by advertisers earlier (so-called third party data, i.e. data provided by independent third party vendors). In this way, advertisements can be accurately targeted as early as the initial stage of the campaign, without a long learning period, therefore the target group of the advertiser becomes available almost right away.
Unfortunately, at the time of the publication of this article, there are still no widely available domestic data, but smaller segments are already available for advertisers. These segments can be searched within the targeting setting of DSPs on a platform similar to a marketplace, where we can search for customised cookie segments belonging to the target group of the advertiser. Based on current experience, the use of pre-defined segments is not suitable for building reach, but could be used for click-to-website purposes in the case of a very narrow target group. The main problem is that we do not know the logic behind the cookies getting into the segments. The only thing we know is that the publisher or DMP named them somehow.
Here we have to mention the Google’s solution for DMPs. Less commonly known, but Google also have an own data management platform, but the approach is different than original DMPs. Google collects data for example from past Google searches, page visit histories and generates audience segments to improve bidding and targeting via Google Display Network. In this case, the data is the property of Google (aka their 1st party data). And most importantly, this data can be used constantly, when Google finally blocks cookie tracking in Chrome in 2023!
Back to our topic: Several technological services may be added to the list of the participants introduced above. Just to mention the best-known ones: brand safety, anti-fraud and other technological service providers, with which the operation of campaigns can be further optimised. Lots of them are built-in to the most popular DSPs or have a native integration with them, so 3rd party vendors are becoming less commonly used. Unique vendors now have to find their working method in / and after the post-cookie apocalypse, so there will be significant changes in this market. Stay tuned.
Mito is a full-service agency with a passion for clever things. The Performance division is specialised in improving digital performance and increasing sales with performance media, SEO, CRO and analytics solutions.
We are proud to have companies like Decathlon, Wizz Air, Cetelem Bank, Office Shoes and Tungsram among our strategic partners.