Google Ads Now Shows You Exactly What Its Recommendations Achieved

Google Ads has quietly rolled out a new “Results” tab inside the Recommendations panel, finally letting advertisers see whether the platform’s automated advice actually moved the needle on conversions, cost per acquisition, or return on ad spend.

For years, Google has pushed machine-learning suggestions—raise budgets, switch to Smart Bidding, add broad-match keywords—but the only feedback loop was a vague “optimization score” that always seemed to climb. Marketers were left guessing whether the score translated into real money or just a warmer glow inside the interface.

That blind spot is now shrinking. The Results tab attributes post-implementation changes to each recommendation, line by line. If you accepted a bid-strategy shift last month, the tab will show how many incremental conversions arrived (or didn’t) and what happened to average CPA. The data is retroactive to the day the change was applied, so you can inspect historical performance without building a custom experiment.

How the new Results tab works

The feature lives inside Tools & Settings > Recommendations > Results. Each row lists a single accepted suggestion, the date it went live, and three core metrics:

• Incremental conversions – estimated lift versus a control group
• Absolute change in CPA or ROAS – not just percentage deltas
• Confidence interval – a range Google assigns to show statistical reliability

Google builds the control group by withholding a small, random slice of traffic that never sees the change. The comparison runs for at least seven days or until the account reaches 100 conversions, whichever is longer. If the interval crosses zero (for example, –5 to +12 conversions), the platform labels the result “inconclusive” rather than green-washing it as a win.

Advertisers can click into any row to view day-by-day charts or export the data as a CSV for deeper modeling. Agencies that manage dozens of accounts can also access a manager-level dashboard that aggregates wins, losses, and ties across their book of business.

Why this matters for European marketers

European accounts often run under tighter privacy constraints: limited cookie windows, consent-mode conversion gaps, and smaller data volumes post-GDPR. Those conditions make it harder to run large-scale A/B tests, so many advertisers simply accept Google’s suggestions and hope. The Results tab offers a lightweight way to validate performance without extra tagging or privacy headaches.

Smaller budgets benefit as well. A Copenhagen retailer spending €3,000 a month can now see whether a €300 budget bump recommended by Google actually returned incremental sales, something that previously required third-party attribution software or a data-science consultant.

Equally important, the tab surfaces negative outcomes. If Smart Bidding pushed CPA up 18 % while conversions stayed flat, the interface flags the recommendation red. That transparency helps marketers build an internal playbook of what to avoid in future quarters.

What sceptics still want to see

Google’s incentives haven’t changed: every recommendation that increases spend or broadens match types grows the company’s own revenue. Critics worry the control methodology could still tilt toward favourable results. Three common questions:

1. Is the holdout traffic truly random, or does Google cherry-pick low-value auctions?
2. Are conversions double-counted when they occur across Search, Shopping, and YouTube within the same account?
3. Will Google publish a white paper detailing the statistical model, or keep it a black box?

Google spokespeople told Search Engine Land that third-party auditors reviewed the methodology and found it “aligned with marketing-mix best practices,” but the company has not released the full technical paper. Until then, seasoned paid-media managers say they will treat the Results tab as directional truth rather than gospel.

Another limitation: the tab only covers recommendations that can be isolated—mainly budget and bidding shifts. Structural advice like “add responsive search ads” or “remove redundant keywords” still lacks post-implementation metrics because they can’t be cleanly split-traffic tested.

Action plan for advertisers

You don’t need to overhaul your workflow to take advantage of the new data. Start with these four steps:

1. Accept only one recommendation per campaign per week. Overlapping changes muddy attribution.
2. Let the test run the full 14-day window before judging; shorter periods often show volatile CPA.
3. Export Results data monthly into a Google Sheet and tag each row with your own internal hypothesis (e.g., “expected CPA drop 10 %”). Over time you’ll build a private knowledge base.
4. Use negative results as leverage in quarterly business reviews. Finance teams love seeing saved spend that would have been wasted.

For agencies, create a simple traffic-light report: green for recommendations that beat target CPA, red for losers, yellow for inconclusive. Clients quickly grasp which automation they should fund and which to ignore.

Bottom line

Google Ads is shifting from