Meta Ads Auction · Total Value, Quality & Relevance

Every time a user opens Facebook, Instagram, Messenger, or the Meta Audience Network, an auction runs for each available ad placement. Thousands of ad sets may be eligible to appear — the auction determines which ad wins each placement for that specific user at that specific moment.

Meta's auction is not primarily demand-driven (users searching for something) — it is supply-driven. Meta has users' attention and must allocate it across eligible advertisers. The system balances two goals: maximising advertiser value (showing the ad most likely to achieve the advertiser's objective) and maximising user experience (not showing ads that annoy or disengage users).

Meta determines the winning ad using a "total value" score:

Total Value = Advertiser Bid × Estimated Action Rate × Ad Quality

The highest total value wins the placement. This means a higher bid alone does not guarantee winning — an ad with exceptional estimated action rates and quality can beat a higher-bidding competitor.

Advertiser bid

The maximum you are willing to pay for the objective action (click, conversion, impression, etc.). Can be a manual cap or automated (Lowest Cost bid strategy lets Meta set the bid to maximise results within budget).

Estimated action rate

Meta's prediction of how likely a specific user is to take the specific action the advertiser's campaign is optimised for (click, purchase, lead form submission, video view, etc.) if shown this ad. This is the most complex component — it incorporates the user's historical behaviour, the advertiser's historical performance with similar audiences, and the ad creative's predicted performance.

Ad quality

An assessment of ad creative quality based on feedback from users who have seen the ad and broader creative quality signals. Ads that receive high engagement, low negative feedback (hide, report), and strong post-click experience score higher. Low ad quality — many users hiding the ad, reporting it as spam, or leaving the landing page immediately — reduces total value regardless of bid.

Estimated action rates are Meta's predictive models — one of the most sophisticated components of the auction system. For each potential ad impression, Meta predicts: how likely is this specific user to click this ad, watch this video, submit this form, or make this purchase?

These predictions are based on:

• User behaviour history. Has this user clicked ads from similar advertisers? Have they bought products in this category? Have they engaged with similar creative formats?
• Advertiser history. What conversion rate has this advertiser historically achieved with similar audiences? What CTR does this creative tend to generate?
• Ad creative characteristics. Video ads, image composition, text length, call-to-action — Meta's models learn which creative characteristics predict higher action rates for different user segments.
• Contextual signals. Time of day, day of week, device type, network connection speed — all affect likelihood of various action types.

The practical implication: providing Meta with more conversion data (through the pixel, Conversions API, and conversion tracking) improves its ability to predict action rates accurately — which directly improves campaign performance. This is why the "learning phase" (the period after campaign launch when Meta collects data) is critical, and why disrupting it with frequent changes degrades performance.

Highest Volume (Lowest Cost) is Meta's recommended default for most advertisers — it lets Meta's AI optimise bids dynamically based on real-time auction conditions. Manual bid caps are appropriate only when you have specific profitability constraints that require hard cost limits, at the cost of potentially reduced delivery.

Meta provides three ad relevance diagnostics at the ad level (not the campaign level) — available in Ads Manager under the Columns selector. Each is rated Above average, Average, or Below average relative to competing ads for the same audience.

Relevance diagnostics diagnose the component driving poor performance — allowing targeted improvement. Below-average Quality Ranking suggests creative issues; below-average Conversion Rate Ranking suggests the problem is post-click (landing page, offer, pricing) rather than the ad itself.

Authentic Sources