Micro Conversions: Feeding Bidding Smaller Signals

A micro conversion is a smaller, intent-revealing action a user takes on the way to your real goal: add to cart, a pricing page view, a demo video play, a started checkout. In an ad account its main job is not reporting; it is giving the bidding algorithm more frequent and earlier signal. Below roughly 30 macro conversions per month, Smart Bidding does not see enough examples to learn, and micro conversions are usually the most practical lever available. Pick the wrong one, though, and budget flows toward people who will never buy.

What is a micro conversion, and how does it differ from a macro conversion?

A micro conversion is a user action that produces no revenue by itself but measurably raises the probability of a purchase. A macro conversion is the actual business goal: a sale, a qualified lead, a subscription. The difference is role, not size. Macro conversions define success; micro conversions define the path to it. The test is simple: do sessions containing that action convert clearly better than sessions without it?

• Ecommerce: add to cart, begin checkout, variant or size selection, wishlist add, shipping calculator use.
• Lead generation and B2B: pricing page view, watching a demo video to the end, opening the quote form, downloading a technical document.
• Services and local business: phone number click, directions request, opening the booking calendar, tapping a WhatsApp link.
• Not micro conversions: page refreshes, cookie banner clicks, 10 second sessions, autoplaying video views and other actions that carry no intent.

Which micro conversions deserve to become bidding signals?

A micro conversion worth bidding on meets four conditions at once: it carries intent, it happens often enough, it is measured reliably, and its correlation with the macro conversion is provable. In practice the working thresholds are at least 100 events per month at account level and a conversion rate at least 3x the site average. Events that fail this filter can stay in reports, but they should not enter bidding decisions.

1. Build a shortlist: sort the 20 most frequent events of the last 90 days in the GA4 events report and strip out technical events.
2. Measure correlation: for each candidate, compare the purchase rate of sessions containing the event with the site average, and drop anything under 3x.
3. Confirm volume: remove candidates below 100 events per month or below roughly 10 per week inside a single campaign.
4. Start with one candidate: adding two micro signals at the same time makes it impossible to tell which one moved performance.

How do you mark a micro conversion as a GA4 key event?

In GA4 the old label 'conversion' is now called a key event, and marking an event makes it importable into Google Ads. You flip a single toggle in the events list under Admin, then import the goal in the linked Google Ads account. The critical distinction: marking a key event in GA4 does not automatically make it a primary conversion in Google Ads. You choose that role during import.

1. In GA4 go to Admin and open Events, then turn on the 'mark as key event' toggle for your micro event. If the event is missing, send it first and wait for data to arrive.
2. Narrow the event when needed: define a dedicated event for the pricing page instead of using all page views, otherwise the signal drowns in noise.
3. In Google Ads open Goals and Conversions, import the GA4 key event, and keep its conversion window aligned with the macro goal.
4. Set the imported micro goal to secondary, so the primary list holds revenue producing goals only.
5. If you assign a value, keep it fixed and consistent. Giving an add to cart the value of a purchase makes reported revenue meaningless.

How do micro conversions feed the bidding algorithm?

There are four ways to feed a micro signal into bidding, and their risk order is clear. The safest is importing the goal as a secondary conversion: the data is reported and enters the learning pool without changing the primary optimisation target. Beyond that come probability based values and remarketing audiences. Making a micro event a primary conversion is the riskiest route and is usually a last resort.

• Import as a secondary conversion: the safest path. The goal reaches Google Ads, gets reported and contributes to learning while the primary CPA target stays intact.
• Journey aware bidding: the beta approach spreading through 2026 lets Target CPA in Search learn from goals that are not bid on directly. Worth testing in accounts with multi touch purchase journeys.
• Probability based value: assign the micro event a value equal to 'probability of reaching the macro conversion times average order value'. In an example scenario where 6 percent of add to cart events end in a sale, the event is worth 6 percent of average order value.
• Audience signal only: keep the event out of bidding entirely and use it for remarketing lists and exclusion rules. Lowest measurement risk, slowest payoff.

What are the risks of bidding on micro conversions?

The main risk is goal drift: an algorithm grows whatever you measure, so an account that overweights micro signals multiplies people who add to cart and never buy. The second risk is value inflation, where a padded conversion count makes ROAS look better than it is. The third is measurement fragility, and the fourth is losing comparability with your own history.

• Goal drift: bidding shifts toward inventory that produces cheap micro events, so macro volume stays flat while cost per sale climbs.
• Value inflation: once micro and macro conversions share one column, conversion rate and ROAS both stop meaning anything.
• Measurement fragility: consent choices, browser restrictions and tagging errors cause more loss on micro events than on macro ones.
• Broken history: the moment you change the goal set, the CPA series breaks. If you do not log the change date, later analysis misleads you.

How do you verify that the micro signal is working?

Validation comes down to one rule: you add the micro signal, but you still judge success on the macro outcome. Freeze the macro conversion count, macro cost per conversion and total revenue for the 28 days before the change, then compare the same three numbers afterwards. Because the learning period typically runs one to two weeks, never decide on the first few noisy days.

1. Record the baseline: macro conversion count, macro CPA and revenue for the 28 days before the change.
2. Change one variable: add a single micro goal as secondary and leave budgets and targets untouched that week.
3. Wait out learning: exclude the first 7 to 14 days from judgement, since a temporary CPA rise is normal there.
4. Compare: if macro cost per conversion and total revenue have not improved, the micro signal added nothing.
5. Roll back and log it: remove the ineffective goal from the secondary list and write the change date into your account log.

A bidding algorithm grows whatever you measure. Adding a micro conversion risks quietly changing the account's goal, which is why the scorecard has to stay on macro results.Ads Sensor Team

Before you touch micro conversions, make sure the macro side is solid: Google Ads conversion tracking setup and GA4 ecommerce purchase tracking are prerequisites for this guide. If add to cart is high but purchases are low, the problem may be the page rather than the signal, and landing page conversion rate is the better starting point. Ads Sensor unifies GA4 with Google Ads and Meta data in one panel and tracks the macro impact of signal changes through automatic before and after comparison; you can connect an account from the beta signup.