When Guillaume Grimbert was working on Google’s DV360 product team in 2019, sustainability wasn’t much discussed in programmatic.
When the topic arose, the focus, he noticed, was on the devices being used or the creative – not ad selection.
But ad selection is a complex and carbon-hungry process. According to Scope3, which measures emissions in the digital supply chain, the majority (60%) of carbon generated by programmatic comes from real-time bidding.
So, in 2022, Grimbert co-founded Greenbids – a Paris-based startup that helps intermediaries reduce carbon in the bid stream – with his Google colleague Adrien Delambre (who was part of Google’s internal sustainability resource group) and Jean-Baptiste Pettit, an expert in machine learning and data science who previously worked at Teads.
Greenbids has a custom algorithm that balances media performance with C02 reduction, which Grimbert referred to as “infrastructure optimization.” In other words, rather than trying to replace existing programmatic systems or forcing buyers to change their media plans, Greenbids aims to make the whole shebang more efficient and, therefore, more sustainable.
“By being everywhere,” Grimbert said, “we can create a network effect.”
For buyers
When it’s working with advertisers, Greenbids ingests their programmatic media plans and connects to their ad-buying platforms directly.
To feed its algorithm, Greenbids pulls a year’s worth of a brand’s historical buying data as well as emissions data collected from organizations like Scope3 and GARM on the verticals said advertiser operates in.
Greenbids then creates emissions benchmarks for campaign KPIs, such as brand awareness, attention or viewability.
For example, if an ad isn’t viewable, there’s no point in buying it. Not only would it fail to convert; it would produce unnecessary emissions. “You could think of our algorithm as trying to reduce the waste in the programmatic chain to make sure that we only have relevant impressions,” Grimbert said.
Once a campaign goes live, Greenbids continues to tweak and adapt as data comes in. At the end of the campaign, it provides clients a report that breaks down how much carbon was saved as well as the impact on cost per view and brand awareness.
Since launching the advertiser product 18 months ago, Greenbids, whose buy-side clients including IPG-owned Mediabrands, Dentsu, L’Oréal and French hospitality company Accor Hotel, has saved 11,500 tons of carbon, according to Grimbert.
For sellers
Greenbids takes a similar approach when partnering with publishers.
It works off historical data and integrates its analytics adapter into existing Prebid modules for header bidding to identify which supply-side platforms should get the bid requests.
Instead of blasting out an ad call through as many SSPs as possible in a header bidding auction, during which duplicated bid requests reach the same demand-side platforms, Greenbids has a machine learning prediction model that only sends bids to relevant SSPs. (It then does the same filtration process with those SSPs to make sure they’re only sending requests to relevant DSPs.)
To determine whether an SSP is relevant, Greenbids analyzes criteria, such as geography. If an SSP that doesn’t sell in the US bids for US-based ads, it gets cut.
Publishers that test the seller-focused tool, which launched in September, see a 20% to 35% reduction in carbon on average.
So far, a major UK-based publisher, which Grimbert declined to name, is using the tool as well as France-based Prisma Media, which is on course to save more than 10,000 tons of carbon this year.
The most recent addition to the Greenbids offering is a YouTube optimization product that was in beta for six months across 30 markets and went live in mid-February.
The tool allows advertisers to buy YouTube ads more sustainably by optimizing when ads are served, on what device, the player size and geography.
A few of its existing clients are now testing the tool, including Accor Hotel and New Zealand-based Robert Harris Coffee Roasters. Average carbon reduction has been around 41%, with cost-per-view down roughly 28%.
On the partnership front, Greenbids recently struck a deal with Magnite to get traffic-shaping data and another with Equativ for access to data from its DSP. This data helps make the Greenbids predictive algorithm more efficient at ad selection and supply-path optimization.
It also has direct integrations with DV360 and Microsoft’s Xandr and is working on an integration with The Trade Desk set to go live in Q2.
But beyond hammering away on its product road map and partnership pipeline, Greenbids is also physically growing. It expanded into the UK this year and is planning an eventual move overseas, including into the US.
And with sustainability standards for the advertising industry coming soon (hopefully), Grimbert feels confident about the position Greenbids is taking as an integrator rather than trying to reinvent how programmatic functions.
“If we try to change habits, it’s really tough,” said Grimbert. “But if we can act on [existing] habits, then we can change the industry really fast.”