This article is a WTF explainer, in which we break down media and marketing’s most confusing terms. More from the series →
Demand for custom bidding algorithms is growing among programmatic advertisers who want an alternative to the off-the-shelf algorithms from ad tech vendors that take a one-size-fits-all approach to buying impressions.
Meanwhile, some agencies see alternative algorithms as a way to exert influence over an ad tech industry that has disintermediated them from advertisers.
Like any new trend, there is confusion around what these algorithms are, and how they work. Here’s a primer.
WTF is a custom bidding algorithm?
Also known as a first-party algorithm, it is a custom set of bidding rules that are dynamically generated on a per-campaign basis and designed to deliver outcomes that are aligned to a specific business’s goals. The advertiser “owns” the algorithm as it is specifically designed for them. Usually, these algorithms are owned by ad tech vendors like demand-side platforms and so can’t be finely tuned to an advertiser’s goals. Some advertisers use their own first-party data to power their custom bidding algorithms, which makes it easier to focus on factors like contextual targeting and attention-based metrics that can be more easily traced back to business objectives.
Where can I find these custom bidding algorithms?
Some ad tech vendors like Beeswax won’t build custom algorithms for programmatic advertisers, but its product makes it easy for advertisers to assemble one themselves. For instance, it will offer advice on how to build a custom algorithm and make introductions to companies like Scibids that develop the algorithms. Scibids alongside MediaGamma and Strike Social are among a new wave of ad tech players going to market with variations on the same pitch: They can develop custom algorithms for advertisers and agencies that make programmatic trading less resource-intensive. It’s a timely pitch for the buy side as the financial burden of making programmatic profitable takes its toll on both advertisers and agencies.
“Marketers and agencies want and need the benefits of domain-specific AI capabilities,” said Matt Nash, managing director of Scibids in the U.K. “However, they also recognize that to build the capabilities themselves would involve a commitment to a multiyear project, dedicated resources and investment on an ongoing basis across multiple platforms with no guarantee of a successful outcome.”
All this sounds expensive.
Hiring a developer to build a custom bidding algorithm isn’t as expensive as it was when the data science and analytical expertise needed to do it were scarce. Furthermore, tools like bid modifiers, which allow ad tech vendors to tweak bids without having to duplicate line items on a campaign plan, now exist. With those skills and services more widely available in advertising now, the barrier to entry for businesses that create custom bidding algorithms is easier, therefore they’re cheaper to buy. That’s led to more investment from both large and small advertisers.
Give an example.
Hotel chain Meliã turned to Scibids, for example, as a way to offset the time and money spent by its in-house team of traders manually tweaking programmatic campaigns every time the objectives changed. The custom bidding algorithm the advertiser paid for generated four times more return on ad spend, compared to campaigns with a standard algorithm, while it also led to a 15% lower bounce rate from its site, said the head of the hotel’s programmatic division, Queralt Costa.
What are the drawbacks?
The hype around what custom bidding algorithms could do far exceeds what is technically possible for most advertisers now.
Some ad tech vendors and consultants pitch the algorithms as tools that will eventually evolve to a point where they get so big that advertisers will be able to take them from one agency trading desk to another. It sounds great in theory but ignores the technical ad ops gymnastics needed to make it happen. Agency trading desks are built on different ad tech so an algorithm that works on one may not work on another. There’s also the commitment to nurturing a custom bidding algorithm to consider. Depending on how big it gets, a custom bidding algorithm would need to be trained to treat campaigns for separate parts of a business like customer acquisition and traffic generation differently. That requires a commitment that not all advertisers are willing to give now that the realities of taking parts of media buying in-house are clearer.
Do I need a media agency to buy my programmatic ads if the algorithm does it for me?
Setting up a programmatic campaign, from putting the right tags in place on the right sites to making sure the right data is used, is not for the fainthearted. It can quickly devolve into an operational quagmire if mishandled as the algorithm won’t make the right decisions in auctions if it doesn’t have the right information. Organizing campaigns to be managed by a custom bidding algorithm is just the start, however. In the case of Mindshare, its execs are still needed to look over what a custom bidding algorithm buys, said Alexis Faulkner, head of Mindshare’s digital division, Fast, in the U.K at a Scibids event in London last week.
“The algorithms are doing the heavy lifting in terms of the day-to-day optimization of programmatic campaigns, but it’s taking a lot of man-hours to evaluate the impact of what they buy,” said Faulkner. “We’re just moving to the analytical phase of custom bidding algorithms.”
It won’t all be plain sailing for media agencies if custom bidding algorithms become more widely adopted. There’s a strong likelihood that agencies won’t need as many programmatic traders as they currently employ if advertisers want an algorithm to place more bids, said Amir Malik, a digital marketing expert at Accenture Interactive at the same event. Instead, agencies will need to set their sights on data scientists who can manage the algorithms that buy the media, he said.
More in Media
The Independent’s Blair Tapper & Thomson Reuters’ Josef Najm are trying to break down advertisers’ news blocks
In a live recording during the Digiday Publishing Summit, the news executives called for more nuanced conversations with advertisers around their brand safety concerns.
After investing in one generative AI startup and suing another, the company will let customers create images on its website and an API.
During the Digiday Publishing Summit, execs from companies including Condé Nast, Dotdash Meredith and Thomson Reuters assessed the industry’s readiness.