How Adriva's Hourly Budget Rebalancing Works Across Google and Meta

Demand changes from hour to hour, but most ad accounts get a budget review once a day at most. A budget set on Monday is often out of date by Wednesday. Adriva checks Google and Meta campaigns every hour and moves budget toward the campaigns that are bringing in conversions.

Why fixed daily budgets go out of date

A daily budget is set once, and in most accounts nobody changes it until the next review. It assumes the conditions at 9am will hold until midnight, which is rarely true. Search demand rises and falls during the day, auction competition goes up when competitors switch their campaigns on, and conversion rates change by the hour. In peak windows, some hours convert at two to three timesthe daily average, and off-peak hours can be 40–60% below it. These are documented dayparting patterns, and most accounts do not act on them.

With a daily review, every decision is based on yesterday’s data. Say a Meta prospecting campaign had a very strong Tuesday. You raise its budget on Wednesday morning, but the hours that made it work are already over. Deciding when to shift spend between Google and Meta by hand is a real skill. Doing it every hour across a dozen campaigns on two platforms is not realistic for one person. Adriva was built to do that part.

Demand changes by the hour, so Adriva reviews budgets every hour instead of once a day.

What Adriva checks every hour

Every hour, Adriva pulls delivery and conversion data for each active campaign from the Google Ads and Meta APIs. Account totals are not enough for this. Adriva reads three signals from the data that help predict where the next dollar should go.

Budget pacing

Pacing shows whether a campaign is on track to spend its budget by the end of the day, or is under-delivering and leaving reach unused. It tells Adriva which campaigns can spend more and which are held back by their budget. A campaign pacing at 130% of plan by noon may be catching a real demand spike that is worth funding. It may also be wasting money and need a cap. The other two signals decide which case it is.

Marginal ROAS and CPA

Lifetime averages can mislead. A campaign with a 6x blended ROAS can still be a poor place for your next dollar if it is saturated and the marginal return on extra spend has dropped sharply. Adriva estimates what the next increment of budget is likely to return, which is the slope of the return curve, and ranks campaigns by that number. Budget goes to the campaign where the next dollar is expected to return the most, on either platform.

Intraday demand

The same audience behaves differently at 9am and at 9pm. Adriva tracks how conversion rate and cost change through the day for each campaign on each platform. It then gives a channel more budget in its strong hours and less in its weak hours, so the budget is not spread evenly across all twenty-four.

These estimates are only as good as the conversion data behind them. That is why Adriva relies on the measurement setup it creates when you connect an account. For the details, see how Adriva connects Google Analytics and the pixels. More accurate conversion data gives more accurate marginal-return estimates.

The hourly rebalancing steps

Rebalancing runs as a fixed sequence of steps, with a check at each one. The same sequence runs every hour.

1. Read. Pull the latest pacing, conversion and cost data from both platforms for every active campaign.
2. Rank. Estimate marginal ROAS or CPA for each campaign and sort the campaigns by where the next dollar is likely to return the most right now.
3. Propose. Find the smallest change that improves the account as a whole. Take a small amount of budget from a saturated or off-peak campaign and give it to one that has room to spend and a strong marginal return.
4. Check guardrails. Confirm the change stays inside caps, floors, step-size limits and learning-phase thresholds. Nothing reaches the account before this check passes.
5. Apply or hold. If the change passes every check and confidence is high enough, apply it through the platform API. If not, hold. In most hours, holding is the right decision.
6. Log. Record the change, the signals behind it and the projected impact, so you can see what moved and why.

Which signal leads to which action

Each signal pattern leads to a specific response. The rules are simple on purpose, so an operator can look at any decision and see why it was made.

Guardrails and the learning phase

Hourly budget changes are only safe if they respect how the platforms learn. Google and Meta both run a learning period while their models collect enough conversions to deliver steadily, and large edits restart it. If rebalancing ignored this, a small gain from better allocation would cost a large drop in delivery. Adriva’s guardrails exist to prevent that.

On Meta, an ad set needs roughly 50 optimization events per week to stabilize, and significant edits send it back into learning. On Google, budget changes above about 20% tend to trigger a reset. Adriva treats both numbers as hard limits:

• Small steps. Each change stays inside the thresholds the platform treats as safe, so no single hourly move restarts learning.
• Holds during learning. Campaigns and ad sets still in active learning are mostly left alone until they stabilize. Adriva moves extra spend into other campaigns instead.
• Budget changes only. Rebalancing moves budget and nothing else. It does not switch bid strategies, change conversion actions or make structural edits that would reset learning on their own.
• Caps and floors. You set the highest budget a campaign can reach and the lowest it can drop to. Adriva only works inside those limits.

The gain from hourly rebalancing comes from many small, well-timed changes that a daily review would miss. Large moves are not needed. Checking every hour lets Adriva time each change better, and the guardrails stop those changes from disrupting the account. We explain the bidding layer below this in a separate post on automated bidding strategies.

What the operator sees and controls

You should be able to check any change an automated system makes. In Adriva, every rebalancing decision is visible and logged, and the operator decides what Adriva is allowed to change.

The log shows which campaign got more budget, which campaign gave it up, the signals behind the decision and the projected impact. It uses the same format as the rest of the product: “shift this into prospecting, here is the expected return”. You can trace every change after it happens.

You also set the rules in advance. Caps and floors limit each campaign’s budget. You can mark a campaign as no-touch, so Adriva reports on it but never changes its budget. You can leave rebalancing fully automatic or switch to recommend-only mode. In that mode Adriva shows the same moves for one-click approval and does not apply them itself. You decide how much Adriva can do on its own, and you can turn automatic changes off again in one click.

Adriva handles the hourly budget checks and changes. You keep the strategy and set the limits.

Pacing checks, marginal-return estimates and hour-by-hour timing have to run all day on two platforms. The work is necessary and repetitive. Adriva does it and shows you each decision, inside the rules you set. You still decide the strategy. Book a call to see Adriva run the hourly rebalancing that is not practical to do by hand.