A Cost/Benefit Analysis of taking Programmatic In-house

In-housing is the hot word in the programmatic space at the moment, especially in the brand world. A quick search for “in-house programmatic” on google news and you get a long list of brands such as Pizza Hut, Treatwell, Orange, PepsiCo and more who have or are executing on moving their marketing and ultimately their programmatic spend in-house. Increasingly brands, especially in the direct to consumer (DTC) space are disillusioned with the traditional brand<>media agency model and are looking for a closer, deeper relationship with their customers that increasingly includes cutting out the middleman. This leads to the natural question around if, and why a core part of your business (marketing) should be managed by an external party.

In this post I’m going to dig in a bit more on the costs of taking programmatic in-house. I’ve provided some good background information on the benefits in this article but to summarize they fall into two buckets.

Benefit bucket #1: Control

When running your programmatic advertising in-house you own:

• IP on “how your UA is done” — this is highly strategic and unquantifiable but a core value add
• Data — limiting who you share your data with means it’s less likely to be leveraged by your competitors
• Algorithm and the feedback loop into how its developed
• An unbiased creative testing framework
• Creative rotation (flighting / taking down)

Benefit bucket #2: Transparency

Driving decisions around:

• Attribution — standardization of how attribution is done
• Incrementality — access to clean testing environments
• Fraud — full access to the bundle ID’s for quick blacklisting/whitelisting
• App usage — build your own session graph to understand how many impressions a user has seen each day, which apps they’d used etc
• Supply Source visibility (Publishers & Exchanges)
• Supply path optimization — cut out the middleman

The benefits are clear, it’s all the control / transparency that you want as a buyer but your UA provider would prefer not to give you — this is almost always due to a misalignment of incentives.

Now we’ve summarized the benefits, let’s look at the costs:

Cost #1: Up-front cash

According to the Association of National Advertisers (ANA), in an August 2018 study of US client-side marketers, “cost efficiencies” were named by far the largest primary benefit for having in-house expertise in buying media.

However — according to a survey by ExchangeWire and IPonWeb of 129 media agency professionals worldwide, 72% of them said that the initial set up cost is the most important factor to consider.

The gap between the rationale behind why a marketer might want to in-house marketing IP and the benefits of in-housing is without a doubt the key reason why in-housing is not yet an accepted route for brands/DTC who can justify the ongoing marginal maintenance costs.

The upfront fixed cost to deploy an in-house solution is why it’s so important to have executive level (CEO/CMO) buy-in for in-housing and for it to be seen as a strategic investment rather than a quick win. The initiative needs real time and money investment and, at least initially, can’t be compared to other channels in that if it’s not performing it gets defunded. Ultimately the ongoing marginal cost of operating your programmatic business will most likely be lower than advertising through managed services due to supply path optimization but I don’t think looking at in-housing programmatic as a cost saving exercise alone is the best way to look at it.

Cost #2: Talent

I’m going to suggest an org structure that I think is necessary to in-house your programmatic advertising but there will be more resources needed from legal, finance etc to run this operation

Programmatic Lead: In-housing is the hot term in programmatic for 2019 but that doesn’t mean there’s a plethora of talent available to staff up your organization. Due to the nascent nature of the space most in-house marketers are learning on the job. Most performance marketers don’t have deep experience in programmatic so there’s a steep learning curve there. There’s a few mobile first companies in the US who’ve taken their programmatic in-house, eg: MZ acquired Fractional Media in May 2015 and at its peak there were 30 people operating the platform spending $110m / year. Wish, who at one point were Facebook’s largest advertiser acquired Tapsense to manage their in-house but the roster of app developers who are currently operating mobile performance campaigns at scale is limited under 100 advertisers. Hiring financial analysts / traders to manage these systems may produce some quick wins in terms of their understanding to how a two sided marketplace works, but this profile would need a significant level of training on advertising domain and their salary expectations maybe significantly different than talent from the marketing world.

Data Scientist(s): If you’re leveraging the algorithms of your tech vendor to run campaigns you won’t need a DS. However, if you’re looking to “own your own IP” as a benefit of in-housing you’re most likely interested in building your own algorithms to leverage your first party data. Data Scientists are expensive and hard to find with ad tech domain knowledge but a good data scientist is worth their weight in gold.

Data Analyst: to fully leverage programmatic targeting capabilities it’s crucial that an advertiser can make sense of the data it holds and can make actionable, in-house analytics, MMP data, targeting options etc. Having a strong analyst with the ability to think programmatically is key when planning and understanding campaigns.

Ad Operations: It would be great to “set it and forget it” but it’s unlikely that this will be the case. Many of the technology providers in the in-housing space have “customer success managers” but ultimately if the plan is to bring programmatic in-house you’re going to need to hire someone on the operational side to manage campaigns.

Cost #3: Technology / Creative

Algorithm: The core piece of technology that needs to be built and fundamental the reason that ad targeting companies exist. Although many advertisers bemoan sharing their data with advertising technology vendors, it’s exactly because the advertiser shares this information that the vendor is able to aggregate a huge pool of data to which the vendor can apply it’s algorithms. Leveraging this pool of data it’s possible to build a rich profile of a device (and therefore a user) based on hundreds (or thousands if you’re Facebook / Google) from which you can build lookalike models to find more people like this user .

Without access to other advertisers data, an in-house solution needs to double down on unique first party data from which to build these user profiles. To start with you’d want to leverage the usual generalized data that you might receive from a mobile measurement partner (MMP) such as publisher name, publisher category, geo, exchange, device make/model etc.

Eg: for example for a game you might ingest events into your algorithm such as:

• Game category
• Level
• Level completion time
• Member of a guild (y/n)
• IAP spend behaviours

Eg: for an eCommerce app examples might be:

• Items bought
• Frequency shopped
• Items in shopping basket
• Items clicked on

However — all these events are extremely custom to the app being advertised and so need to be piped into a data stream from which your data scientist mine.

Technology: According to Digiday the cost to partners with a DSP (demand side platform) technology to bring programmatic spend in-house is “typically 10–15 percent for self-serve solutions”. Although one objective of in-housing might supply path optimization to reduce costs there are costs to accessing this supply leveraging a 3rd party technology due to the infrastructure costs of “sniffing” the bid requests and maintaining and developing the technology.

Creative: On mobile specifically, many companies utilise the technology vendors they work with to build specific creative for their platforms. In-housing can result in additional overhead required to build this creative for the in-house platform

Cost #4: Opportunity

There’s an opportunity cost to any decision, so it’s crucial that the benefits of the strategic initiative outweigh the cost any other action you could take to grow your business. For some this cost may be too high and they should focus on their core business or other marketing activities but for others who see the strategic investment opportunity it that in the long term can make your business.

From Eric Seufert’s interview on the How Things Grow Podcast:

“that thing (in-housing programmatic) is basically like a cash printing machine, because you’re seeing all this scale in places that other people can’t see, and they don’t have access to…it’s really powerful but, there’s a huge amount of upfront cost that you need to be comfortable with before you start down the path. And I think the way that people mess that up is that they try to do it cheaply, and they try to cut corners”