Automated bidding is when the ad platform sets your bid for each auction, using a model that predicts how likely that impression or click is to lead to the result you've chosen. AI-run campaigns go further: the platform also chooses the audiences, the placements and the combinations of your creative. In B2B, hand over the decisions inside each auction, which the platform can make faster than a person. Keep the decisions that need knowledge of your business: what counts as a good result, who to reach and exclude, what the ads say, and how you measure whether the spend worked.

Automated systems can adjust a bid for every auction using signals no person could process in time, such as device, location, time of day and browsing behaviour. The risk in B2B is that they find the cheapest way to get the result you've asked for, and the result that's easy to give them, such as a form fill, is often a long way from revenue.

What Automated Bidding Does

In an ad auction, your bid decides whether your ad shows and what you pay. With manual bidding, a person sets bids for keywords or audiences. With automated bidding, the platform sets a bid for each auction based on its prediction of how likely that impression or click is to lead to your chosen conversion, and what that conversion is worth. Common strategies aim for a target cost per conversion, a target return on ad spend, or the most conversions within a budget. Automated bidding is a form of predictive AI: it ranks each opportunity by how likely it is to produce the result.

AI-run campaign types hand over more. You supply a budget, a goal, creative assets and some audience signals, and the platform decides where the ads run, who sees them and which combination of headline, image and copy each person gets. For example, Google's Performance Max and Meta's Advantage+ campaigns work this way.

Why B2B Is Harder for Automation

• Fewer conversions: a B2B advertiser may get a few dozen qualified leads a month, where an online retailer gets thousands of purchases. The model has less to learn from, and results swing more from week to week.
• Long buying cycles: a deal can close months after the first click, long after the platform has stopped looking for a result.
• Revenue is recorded somewhere else: pipeline and closed deals sit in your CRM, so the platform can't see them unless you send them back.
• A group decides: the person who clicks and fills in the form often isn't the person who approves the purchase. A B2B buying committee typically has 6 to 10 people in it.

The Platform Optimises to What You Give It

An automated bidding system does what it's told. If the conversion you've set is a form fill, the system looks for people who fill in forms at the lowest cost. Students, job seekers, competitors and people collecting free content all fill in forms. Cost per lead falls, lead volume rises, and the share of leads that become pipeline falls with it.

The system also makes trade-offs you didn't ask for. Ads for jobs in science, technology, engineering and maths were shown to fewer women than men, even though the advertiser set them up to be gender neutral, according to Anja Lambrecht and Catherine Tucker in Management Science (2019). Younger women were more expensive to reach, so an algorithm built to minimise cost showed the ads to more men. In B2B, the same logic can push spend toward the cheapest impressions, which are rarely the senior decision makers you need to reach.

Automation Can Take Credit for Sales That Would Have Happened Anyway

Brand-keyword search ads had no measurable short-term benefit in large field experiments at eBay, published by Thomas Blake, Chris Nosko and Steven Tadelis in Econometrica (2015). For non-brand keywords, most of the spend went on frequent buyers whose purchases weren't influenced by the ads, so average returns were negative.

A system trained to find conversions will find the people most likely to convert, and many of them are already on their way to buying. Its reported cost per conversion can look excellent while the campaign adds very little. The way to find out is to test what happens without the ads, using a holdout audience or a test in a few regions. Attribution vs incrementality covers how to set that test up.

What to Hand Over and What to Keep

How to Set Up Automated Bidding in B2B

1. Send the platform a conversion as close to revenue as you can. Import qualified opportunities or pipeline from your CRM, rather than relying on form fills alone.
2. Give each conversion a value. For example, value a sales-accepted opportunity far higher than a content download, so the system learns which result matters to the business.
3. Give the model enough to learn from. Combine small campaigns that share a goal, rather than splitting the budget across many campaigns that each get a handful of conversions.
4. Set exclusions. Exclude existing customers, competitors, job seekers and regions you don't sell to.
5. Change one thing at a time. Large changes to budgets or targets send most systems back into a learning period, and results are unstable until it ends.
6. Test incrementality. Hold out a region or audience, and compare pipeline with and without the ads.

FAQ

Does automated bidding work for B2B? It can, when it has a conversion signal close to revenue and enough of them. With only form fills as the signal, it tends to find cheap leads that don't become pipeline.

How many conversions does automated bidding need? Each platform publishes its own guidance, and it changes. The fewer conversions a campaign gets, the less reliable the model. If qualified opportunities are too rare, use the earliest stage that still predicts pipeline, such as a sales-accepted lead, rather than any form fill.

Should B2B advertisers use AI-run campaign types? Test them against your current campaigns with a holdout. Keep control of exclusions, brand safety and the conversion signal, and judge the results on pipeline rather than platform-reported conversions.

What's the main risk of automated bidding in B2B? Optimising to a conversion that doesn't lead to revenue. Cost per lead falls, lead volume rises, and pipeline stays flat.

Hand Over the Auction, Keep the Strategy

The platform can decide what to bid in a single auction better than a person can. It can't decide who your customers are, what a good result is worth to the business, or whether the spend made a difference. Those decisions stay with the marketer, whatever the campaign type. B2B Marketing Fundamentals covers the strategy decisions that come before any campaign, and ends with a one-page growth strategy for your own business.

Sources

• Thomas Blake, Chris Nosko and Steven Tadelis, "Consumer Heterogeneity and Paid Search Effectiveness: A Large-Scale Field Experiment", Econometrica, 2015. Source of the findings on brand-keyword and non-brand search ads at eBay. nber.org
• Anja Lambrecht and Catherine Tucker, "Algorithmic Bias? An Empirical Study of Apparent Gender-Based Discrimination in the Display of STEM Career Ads", Management Science, 2019. Source of the finding that cost-optimised ad delivery showed a gender-neutral ad to fewer women. dspace.mit.edu
• Gartner, "The B2B Buying Journey". Source of the finding that the typical buying group for a complex B2B solution involves 6 to 10 decision makers. gartner.com