The budget value-based bidding needs to learn

Why there is a floor at all

Google's Smart Bidding and Meta's delivery both set a bid for every auction from what they have learned about which clicks turn into conversions. Learning needs examples. Below the floor neither sees enough of them to tell one kind of lead from another, so the bids stay close to guesses.

This is the part that catches people out, because nothing fails visibly. The campaign runs, leads arrive, the reports fill in. What does not happen is improvement, and three months later the conclusion drawn is usually that value-based bidding does not work, when what actually happened is that it never had enough to work with.

The arithmetic, in full

Google Ads: your cost per lead times 30 is the monthly spend per bid strategy. Divided by 30.44, the days in an average month, it is the daily budget.

Meta: your cost per result times 50 is the weekly spend per ad set. Divided by 7, it is the daily budget.

There is nothing else in either, which is deliberate: a budget recommendation you cannot check on the back of an envelope is a recommendation you should not act on.

What the arithmetic cannot do is predict. It uses today's cost per lead, and your cost per lead changes as soon as the bidding changes, which is the reason for doing any of this. Read the result as the floor to start above.

If the floor is out of reach

Raising the budget is the obvious answer and usually the worst one. Three of these four cost nothing.

• Consolidate before you spend

  Google Ads. Five campaigns at eight conversions a month are five strategies that never learn. The same spend behind one bid strategy, or a portfolio strategy across them, is forty, which does.

  Meta. Five ad sets at ten results a week each never leave learning. The same spend in one or two ad sets gets there.

• Optimise on an event that happens often enough

  Google Ads. Bid on the conversion action that clears the floor. A later stage can come once volume allows.

  Meta. On instant forms, if the Conversion Leads stage is too rare to reach fifty a week, map an earlier stage until volume grows.

• Widen the targeting

  Google Ads. Tight match types and audiences starve the strategy of examples. Broadening costs nothing and is easier to undo than a budget increase.

  Meta. Narrow audiences do the same to an ad set. A broader audience gives delivery room to find the results it needs.

• Count the conversions you already have

  Google Ads. Leads reaching your CRM without a Google click ID count toward nothing, because Google cannot attribute them.

  Meta. Leads sent without an email, a phone number or Meta's cookies cannot be matched, so they count toward nothing either.

Volume is necessary, not sufficient

Clearing the floor gets the bidding enough examples. It does not give it anything to sort them by. Value-based bidding works on the difference between what one lead is worth and what the next one is worth, so a feed where every lead carries the same value makes Maximize conversion value on Google, or value optimisation on Meta, no different from counting leads: the same ranking, the same spend, the same leads.

Both halves are checkable against your own data. The diagnostic reads a CRM export in your browser and reports your real volume, the spread between your leads and whether the platform can match them at all. The assessment is ten questions and takes two minutes if you would rather not upload anything yet.

Questions

Why 30 a month on Google and 50 a week on Meta?

They are each platform's own published guidance. Google gives 30 conversions a month as the point where value-based Smart Bidding has enough examples to work with. Meta's ad sets leave the learning phase after about 50 optimisation events in seven days, and Meta offers value optimisation on leads at the same pace, about 100 valued events in 14 days. Below either, the bidding is not wrong so much as under-informed.

Does spending this much guarantee the conversions?

No, and the calculator does not claim it. It multiplies your current cost per lead by the floor. Your cost per lead moves once bidding changes, which is the whole point of changing it, so treat the result as the floor to start above rather than a forecast.

Is the threshold per campaign or per account?

Neither, exactly. On Google it is per bid strategy, so a campaign with its own strategy needs its own volume. On Meta it is per ad set. Either way, splitting one healthy campaign into five themed ones is the commonest way an account falls below the floor without spending a penny less.

What counts as a conversion for this?

Whatever the bidding optimises toward. On Google, the conversion actions the campaign bids on, including imported offline lead values once Google attributes them. On Meta, the ad set's optimisation event, such as a valued lead, or a pipeline stage for Conversion Leads on instant forms, once Meta matches it. On both, a lead the platform cannot match back to an ad counts toward nothing, which is why matching matters as much as spend.

What if I cannot afford the floor?

Consolidate before you spend. Fewer campaigns or ad sets sharing the same spend pool it into one learning set. Broader targeting helps too, and on instant forms so does mapping an earlier pipeline stage that happens more often. Raising the budget is the last option, not the first.

Does this apply to tROAS and to Meta's value optimisation?

Yes, and more so. Target ROAS and Maximize conversion value on Google, and value optimisation on Meta, learn from the spread between one conversion's value and the next, so they need both the volume and enough variety in the values. Volume without variation gives them nothing to sort.