In this three part series Timothy Whitfield, director, technical operations at GroupM shares his adtech predictions for 2020. Read part one here.

Following on from the previous article about the prediction of AdTech in the year 2020; here is part two of that same series.

The article is founded on the premises the triad of pressures on the advertising industry remain (a) increased transparency (b) increased efficiency and (c) increased performance.

Data Management Platforms

A DMP is the pixel tracking technology that is deployed on the advertiser’s website in order to segment customers (visitors) into various demographic, sociographic, interest segments. This segmentation is useful to understand who visits the site, what they purchase and how to find more visitors like them on the internet.

DMPs (like Adobe, Krux, Oracle etc…) that can house advertisers first party data will continue to grow strongly over the next few years. Advertisers will “double down” on these technologies and this will enable the media agencies and creative agencies to build richer a more meaningful advertising strategies. However, this group of companies is not without it’s problems.

New from the Which-50 Digital Intelligence Unit : CMO of Tomorrow edition

CMO of Tomorrow Study Digital Intelligence Unit members ADMA and Oracle Marketing Cloud

  • Very complex deployments – Deploying a DMP throughout an advertisers site is not a very easy task. It sounds easy enough in the sales pitch however when rubber hits road then that’s when problems arise. Advertisers are building more complexity into their site design so that it works with different types of devices. This means more complexity at time of deployment and more data points that need to be carefully collected. Maintaining a DMP is a full time task that requires the IT team, the product team and the marketing team as well as the creative and media agencies. Don’t see DMP deployment as “set and forget”. Instead, see it as the life blood that feeds a successful campaign.
  • Prices are astronomical – DMP prices are unfathomable. Remember that technically a DMP is really a spreadsheet on steroids. It has rows of people that visit a site and columns of interesting facts about these users. What will most likely happen between now and 2020 is that light weight DMPs will be created. Companies that realise that there is good margin to be made here to simplify the deployment, servicing and reporting from DMPs.
  • Data quality and quantity – Having a DMP without any data in it is like having a warehouse without any stock. Just empty shelves. Therefore the first task will be for DMPs to ensure that they have vast quantities of fresh data in the warehouse. Advertisers will start investing in in-house data scientists that will have personal KPIs measured in data quantity and quality.In summary: Data Management Platforms – DMPs will reduce their price and streamline their currently over complicated deployment processes.

Mobile – Location Services

Hyper Local Mobile Targeting is the ability to know where a potential client is currently geographically located and target them with a piece of relevant advertising. There are many pitfalls with this technology and the accuracy of IP based technology is very dubious. Here are some of the things that may happen with the push towards 2020.

  • Standardization of services – There are many companies that offer location based services (, Bliss, Snakk etc..). There is no standardization on how these technologies work. There are no published best practices. There is no regulation from the IAB or any other governing body. Remember that one of the main forces at the moment is increased transparency. Therefore, we can expect more companies to be more forthcoming with information on how their technology work ‘under the hood’. This doesn’t mean that they will be giving away their IP, it just means that they will be willing to share how they find out when somebody is (a) at home; (b) at work (c) standing in front of a shop. etc…
  • Migration to beacon – We can expect to see a slow down of Mobile Services that are IP based and a rise of location based services that are beacon based. The problems with IP based location services is that the accuracy is poor. Meanwhile, beacon’s are small “puck like” devices that are placed in bricks and mortar locations. These beacons are very accurate as their Lat/Long addresses are known and they use Bluetooth technology which only has a few meters range. A lower broadcast range actually means a higher accuracy as it means that the user was standing closer to the device in order for their location to be known. By the year 2020, there will be mobile based beacon services that are able to be used in DSPs in the same way that people can select (a) inventory or (b) data segments.
  • Deployment of SDKs – In order for the beacon services to work then mobile phone developers will need to deploy a mobile location service Software Development Kit (SDK) into their application. There are several problems with deploying these SDKs. Firstly, they increase the overall size of the application, they increase risk because it’s hard to know exactly what the SDK owner does with the data and lastly the SDK, like any piece of software, needs to be updated from time-to-time. However, all of these micro challenges need to be overcome as mobile location based SDKs need to be installed into a far greater range of mobile apps in order for location based targeting to work optimally.

In Summary: Mobile – Location Services – Programmatic traders will be able to select their Mobile Geo Location service provider in the same way that they currently select (a) programmatic inventory or (b) programmatic data segments.

Mobile – Creative Services

Mobile creative services are the types of companies that produce high impact creative formats (Celtra, Playground XYZ, Sizmek’s creative suite, Innovid to name just a few). These companies are growing well because of the fragmentation of media consumption; however things like responsive design creative; programmatic rich media creative will start eating into their healthy profit margins. The main three pressures on this area of the ecosystem are as follows;

  • Migration to “tag-less” deployment – Some of the mobile creative services require the publisher to deploy more JavaScript onto their site in order for the creative services to work. More JavaScript is bad for two main reasons; (a) for the consumer it makes the site slower to load and (b) for the web master it’s another piece of software that could go wrong. Some of the more advanced mobile creative service companies have found ways to develop their high impact creative into standard 300×250 ad-units which are available programmatically and don’t need any more JavaScript to be deployed in order to make it work.
  • Increase in Dynamic Creative – When looking at ad-server data over a large period only about < 2 per cent of all creative are dynamic. At a recent IAB summit it was stated that about 70 per cent of all mobile ads are still static JPG files. These statistics are disgraceful. Mobile creative services will become more and more dynamic into the future. This hardly revolutionary thinking however it just needs to be underlined how important it is to migrate to more attractive/engaging creative.
  • Price needs to come down – Mobile creative can still be very expensive. It can cost upwards of 25 per cent of the media fee in order to host dynamic rich media creative on a mobile device. As apposed to standard ad-serving which can cost only circa 1 per cent of a mobile inventory fee. It is more complicated to host a dynamic/interactive ad, however it’s not 25x more complicated. Prices for this sector will decrease drastically over the coming years. This is exactly what we saw with standard third party ad-serving and it’s what we will see with Mobile Creative Services.

In Summary: Mobile – Creative Services –  Rich, dynamic and engaging mobile ads will be able to be purchased programmatically, at scale, just like any other advert.

Mobile – Inventory Services

Mobile Inventory services is where a publisher wants to put their mobile web or mobile In App inventory online and make it available programmatically.

There are a number of challenges in this area of the market. These are;

  • Unique Inventory Pools – There are so few unique pieces of inventory left in the open market. For instance: a well known weather site can be purchased 34 different ways. Which one of the sellers of this site will offer the best inventory for the advertiser? Which one of these resellers is sitting higher/lower in the waterfall? Mobile Inventory services need to be more transparent about the inventory that they have to offer, they need to explain if the inventory is unique, any special features like bid bias, first look.
  • Mobile SSP Specialists – Integrating mobile inventory, especially in-app, inventory is harder than surfacing general display inventory. There is extra complexity because there are hundreds of thousands of apps and each app is different. These differences make it harder to deploy ads and easier for something to go wrong. There will be Mobile SSP specialists in the same way that we currently have video specialists. These specialists will surface their inventory into both display and video DSPs at scale.
  • Multiple SDK deployment – One large challenge for this sector will be the deployment of multiple SDKs. The pressure from the advertisers and agencies will want more measurement and better targeting which means more SDKs being deployed. However, publishers will want fewer pieces of foreign code in their apps. Therefore the Mobile Location Service partners will need to become experts in deploying SDKs into their apps.

In Summary: Mobile – Inventory Services – Mobile Inventory services will become experts and specialists and experts in integrating with other mobile measurement platforms.

Native Advertising Platforms

Native advertising (like Plista, Outbrain, Taboola) is whereby there is some recommended content at the bottom of a normal editorial style web site. Native is still relatively new and will continue to rise for the next few years. Some key things to lookout for in native include;

  • In-click versus out-click – Some native platforms have already adopted the In-Click methodology and seen huge success. In-Click is when a user clicks on a piece of native content and the user is NOT directed to a new website but instead is redirected to an page on the current website. This is a much cleaner and more “trustworthy” experience for the user. This results in a much higher engagement rate with the user and a significantly increased performance for the media.
  • Integration with CMS – A CMS is a Content Management System. Large CMS systems like WordPress allow the user to create full customised web sites. However, one feature in native platforms that seems like it will grow is the integration of native ads into a CMS system. Each web site that native ads appear on has a different look and feel. For example each site will have a different set of fonts, colours, layout etc… By integrating CMS functionality it means that web sites can have native content that looks and feels like their own web site, even when they click on the ad and it redirects to another article.
  • Limited Native Inventory – One problem with native that display doesn’t have is limited inventory. Each publisher that has native content needs to undergo an integration process. That process can be time and resource consuming so there isn’t any point deploying a native widget on a site that only gets a few thousand visitors a month. Therefore, native widgets are generally only deployed successfully on larger web sites and generally speaking each publisher will only have one native partner on their site. Native is also affected by ad-blockers which will block the whole native widget from loading. All of these factors combined means that native inventory will remain limited.

In Summary: Native Advertising Platforms – Native platforms will continue to deepen the relationship with publisher content management systems in order to generate far superior engagement and performance rates.

Publisher Ad-Servers (PAS)

Publisher Ad-Servers are the types of technology that a publisher places within their CMS (Content Management System) in order to sell their inventory. The key function for a PAS is to correctly manage the booked inventory. There are only a few big players in this space (DoubleClick for Publishers, SAS, AppNexus (previously Open Ad-Stream)). However, even though there aren’t many players, this space is set to explode in the near future. Some of the exciting things that hopefully will happen here are;

  • Automatic Guaranteed – Automatic Guaranteed is whereby the publisher ad-server “talks” directly to the buying playform and allows the agency to (a) check availability; (b) negotiate rates; (c) book campaigns and (d) upload creative directly into the ad-server. This technology is at the heart of all three pressures on the industry. (1) It offers more transparency as the agency can see rates and avails directly in the system; (2) it has massive efficiency gains as it connects two technologies that are normally separated by half a dozen manual processes and (3) it can offer more performance as the inventory can be purchased very granularity (more targeting granularity usually equates to better performance).
  • Header Bidding – Header Bidding is a small JavaScript file called PreBid.JS and it’s deployed on the publishers web site. The file is a “hack” that allows for the publisher ad-server to surface information about the inventory in the auction. For instance; the publisher ad-server may know that some inventory has a high viewability and therefore be able to sell that at a premium.Let me be very clear on this. Any solution that requires more code, more latency, more JavaScript, more data privacy concerns in order to work is a bad solution. Publishers that are thinking about deploying Header Bidding as a technology should think long and hard before taking that step. It’s risky and may yield no results and in fact just slow down the site.
  • Multi-Format (inc Video) – It makes absolutely no sense that a publisher should deploy technology A for display advertising and technology B for video advertising. DoubleClick for Publishers (DFP) is one of the leading technologies that works well for both display and video. Other Publisher Ad-Servers need to embrace this and also build out the multi-format functionality.

In Summary: Publisher Ad-Servers – Publisher Ad-Servers will reject technologies, like Header Bidding, that requires more code, creates more latency and detracts from the user experience will ultimately fail.

Supply Side Platforms (SSP)

Supply Side Platforms are the technology that maximised the yield of a programmatic ad for a publisher. They are normally deployed within a Publisher Ad-Server and inventory that isn’t sold directly is passed to the SSP for monetisation.

The SSP market is in trouble. If the SSP is part of a full stack solution then they will be protected. This is because inside a full stack they will transparently connect to the DSP and increase targeting granularity and therefore increase performance. SSP’s that deal mostly in video will also be protected because good video inventory is at a premium. However, pure-play display SSP’s are in trouble.

  • Automatic Guaranteed – This technology will move money away from the pure-play SSPs and move them directly to the Yield Optimisation engines that are connected to the Publisher Ad-Servers
  • Header Bidding – This technology will slow down the page load. The literature says that it won’t slow down the page load but that is only based one very specific cases. It gets very technical beyond this point as the discussion devolves to an investigation in parallel versus serial auctions. However, we (the industry) can’t solve problems by throwing more pixels, more JavaScript and more code on publishers pages. Solutions need to be elegant not hacks.
  • Ad-Fraud – Advertisers are demanding higher quality inventory. The war on Brand Safety seems to be fought and won but the war on Ad-Fraud hasn’t really even started. There are many web site which still buy traffic. They purchase “pop-under” traffic and automatically load the web site that contains ads. A good way to check is to look at utility sites like SimilarWeb and look for Referral traffic that is from a known click farm.

Let me be 100 per cent clear on this point; it’s the job of the SSP to remove the Ad-Fraud. If the SSP has given a programmatic ad-tag to a fraudulent publisher then they are in the wrong. There is no debate on this point.

In Summary: Supply Side Platforms – The ‘pure play’ display SSP market will find it very tough. Expect to see significant evolution, consolidation and re-invention in order for them to survive.

This concludes this part of the summary. Topics still to be covered include:

  • Search and Social Platforms (inc. Aggregators, Sentiment, Attribution)
  • Tag Management Solutions (inc. Site Optimization)
  • Out of Home (inc. Programmatic OOH)
  • Video Ad-Servers (inc Programmatic TV and Server Side Ad-Stitching)
  • X-Device Specialists (inc. Atlas, Facebook and Google)
  • Brand Safety & Ad-Fraud (inc. Brand Safe Blocking vs Pre-Bid)
  • Viewability (inc. Benchmarks, IAB standards, and trading on viewability)
  • Artificial Intelligence in Marketing (inc. affect on Re targeting Specialists)
  • Ad-Blocking (Publisher Paywalls, Ad-Blocking Specialists)
  • Personally Identifiable Information (inc. Auditing web sites for data leakage)

This story was first published on LinkedIn Pulse and has been republished here with permission. Read part one here.


Previous post

Instagram passes the half a million advertisers mark

Next post

Customer experience driving online sales growth, says Coles GM

Join the digital transformation discussion and sign up for the Which-50 Irregular Insights newsletter.