How to set a Google Ads budget that won't overspend

You set a £30 daily budget. You open the account on Thursday and yesterday cost £54. Nothing is broken, nobody changed a setting, and the number in the budget field still says £30.

Your Google Ads budget is an average, not a cap. Google can spend up to twice your average daily budget on any single day, and up to 30.4 times it in a month (30.4 is 365 divided by 12). Those two numbers are the actual limits. The daily figure you type in is a target that Google paces against, spending more on days it expects to convert and less on quiet ones. You are only ever billed up to the monthly limit, so a £30 daily budget bills at most £912 a month, whatever any single day looked like.

That gap between the number you set and the number you see is where most budget panic comes from. Here’s how the mechanics actually work, and how to size a budget from your own numbers rather than someone’s blog-post minimum.

How much can Google spend in a day?

Two limits, both derived from the one number you set. Google’s own documentation puts it plainly: your daily spending limit is two times your average daily budget for most campaigns, and your monthly spending limit is 30.4 times it.

So for a £30 average daily budget:

Limit	Calculation	Amount
Daily spending limit	£30 × 2	£60
Monthly spending limit	£30 × 30.4	£912

The £54 Thursday is inside the rules. Google calls this overdelivery, and it compensates elsewhere in the month, so some days you’ll see 20-40% of your average daily budget spent instead. If served costs run past the monthly limit anyway, that’s Google’s problem rather than yours: you aren’t charged above the monthly spending limit.

Two practical consequences. First, never set a daily budget you couldn’t survive at 2x on a bad day, because a bad day is a normal, documented outcome. Second, judge pacing on 30 days, not on 3. A three-day sample of a system that deliberately spends unevenly tells you nothing, and reacting to it by cutting the budget resets the learning you were paying for.

What changed for scheduled campaigns on 1 June 2026

If you run ad schedules, weekdays only or business hours only, the maths under you moved this year.

Google changed how it paces budgets for scheduled campaigns with effect from 1 June 2026. Campaigns now pace toward the full monthly limit of 30.4 times the daily budget, whether or not ads are eligible to run every day. Previously a weekday-only campaign paced across its roughly 22 active days.

The caps haven’t moved. A campaign still won’t exceed 2x the daily budget in a day or 30.4x in a month, and ads still don’t serve on days you’ve switched off. What’s changed is the concentration: the same monthly pot now gets spent across fewer days, so your active days run hotter.

If you set a weekday-only budget before June and haven’t looked since, your effective daily spend on Monday through Friday went up without you touching anything. Two things to check:

• Your daily budget field. If you sized it expecting ~22 active days of spend, you’re now pacing at ~30.4 days of spend squeezed into those 22.
• Whether the schedule earns its keep at all. Schedules made more sense when they saved money. Now they mostly redistribute it. If you’re running one out of habit rather than because leads at 3am genuinely don’t convert, the schedule is costing you reach for no saving.

How do you work out a starting budget?

Ignore the “spend $10 a day” advice. It isn’t wrong so much as unanswerable, because it’s a number with no relationship to your business. Work backwards from what a customer is worth to you instead.

You need four numbers, three of which you already have:

1. Target cost per acquisition. What you can pay for a customer and still be happy. If your gross margin on a first order is £120 and you want half of it, your target CPA is £60.
2. Landing page conversion rate. From your analytics. If you don’t have it, assume something conservative and treat the first month as buying the number.
3. Expected cost per click. Google’s Keyword Planner will give you a range for your terms. Take the top of it.
4. How long you’re prepared to wait for a signal.

Then: CPC ÷ conversion rate = cost per conversion. At a £2.50 CPC and a 4% conversion rate, that’s £62.50 per conversion, just past a £60 target, which is a useful thing to learn before you’ve spent anything.

For the budget itself, aim for enough daily spend to buy roughly one conversion a day. Below that, you’re not gathering data so much as sampling noise, and Google’s bidding has nothing to learn from. At £62.50 a conversion, a £60-a-day budget is honest; £15 a day is a science experiment with a sample size of one per week.

Smart Bidding needs conversion volume to work with, which is the real argument against tiny budgets. It isn’t that £10 a day is morally wrong, it’s that a bidding algorithm optimising on two conversions a week is guessing, and you’re paying for the guess. If your target CPA maths says you can’t afford a conversion a day, that’s a signal to narrow the campaign, with fewer keywords or tighter geography, until the budget is dense enough to teach the system something.

Daily budget or shared budget?

A shared budget is one average daily budget pooled across several campaigns. Google reallocates it: underused budget moves to campaigns that are capped, so you spend the whole pot instead of leaving money in a campaign nobody searched for that day.

Which one you want depends on whether you’d be happy with the reallocation Google chooses.

Shared budgets suit accounts with several campaigns pointed at the same goal, seasonal pushes, or anyone whose real constraint is a single monthly number rather than a per-campaign one. They cut a lot of fiddling as the account grows.

Individual budgets suit campaigns you need to protect. The failure mode of sharing is well known: a £50 shared budget across five campaigns will not spend £10 each. One campaign can take £48 of it and leave the others invisible for the day. If one of those campaigns is your brand terms, or a product line you’ve promised to support, don’t pool it.

The rule of thumb that survives contact with real accounts: pool campaigns you’d be indifferent between, and separate campaigns you’d argue about.

Why did my spend jump when I didn’t change anything?

Assuming you’re inside the 2x and 30.4x limits, the spend went up because something about the auction or your own campaign changed underneath you. In rough order of how often it turns out to be the culprit:

Google shifted pacing within the month. The likeliest and most boring answer. Higher spend today gets paid back with lower spend later, and the monthly total lands where it should. Check month-to-date against your monthly limit before doing anything.

A budget-limited campaign got unblocked. If you raised a budget, paused a competing campaign, or a bid strategy loosened, a campaign that was previously capped starts buying the demand it was turning away. The extra spend was always available, you just weren’t paying for it.

Average CPC rose. Seasonality, a new competitor bidding aggressively, or a quality score slipping on your best keyword. Spend stays flat while clicks fall, or spend rises while you hold click volume. Segment by keyword over 30 days and the offender is usually obvious.

Match types found new ground. Broad match and Smart Bidding will keep discovering queries, and some of them are expensive. The search terms report over the period in question tells you whether you bought something new, and whether you’d want to buy it again.

Someone else in the account made a change. Including an automated rule, a script, or an agent. This is the argument for having a log of who changed what and when, rather than reconstructing it from memory and change history a fortnight later.

Work down that list before touching the budget field. A budget cut reacting to normal pacing is the most common self-inflicted wound in a small account: you cap spend on the days Google identified as your best converting days, then wonder why CPA got worse.

What to check every week

Budget management is mostly a maintenance job, and it’s a good one to hand to an agent because it’s four queries and a comparison, every week, forever.

• Actual spend against the monthly limit, not against the daily number. Month-to-date spend divided by days elapsed, compared with your average daily budget, tells you whether pacing is where you think.
• Campaigns flagged “limited by budget”, which means demand exists that you’re declining to buy. That’s either a budget increase or a targeting problem, and it’s worth knowing which.
• CPC drift. A rising average CPC quietly shrinks the number of clicks your budget buys. Same spend, less traffic, and it doesn’t announce itself.
• Whether anything changed the schedule or the pacing assumptions, which after June is the one to watch on scheduled campaigns.

If you’re running this through an agent, it’s a google_stats call at campaign level over a 30-day window, compared with the budget on each campaign. The useful part isn’t the pull, it’s that the pull happens every week whether or not you remembered. If you’re weighing that up against paying someone to do it, the economics of running your own ads versus an agency come down to exactly this kind of recurring, unglamorous check.

One caution on automation. Budget changes are money-bearing, so the agent that reads your pacing shouldn’t be the agent that quietly triples a daily budget at 2am because a rule fired. Read on a schedule, write on a decision. If you’re wiring an agent into the account for the first time, start with the read tools and only hand over writes once you’ve watched a few weeks of its reasoning.

---

Budget checks are a weekly job, so hand them to something that doesn’t forget. FlyWheel gives your AI agent one MCP surface across Reddit, Google Ads, Meta, and X, with every tool call logged — args, status, latency, actor — and new campaigns shipped paused by default. Get started with FlyWheel.