How do you calculate the floor?
With a chain of four numbers you either have or can estimate:
- Sales per month you are after. The target, not the aspiration.
- Close rate. Out of ten qualified opportunities, how many do you close?
- Landing page conversion. Out of a hundred visits, how many leave their details?
- Cost per click in your sector. Google's keyword planner estimates it.
Sales ÷ close rate = opportunities needed. Opportunities ÷ page conversion = clicks needed. Clicks × cost per click = monthly budget. If the result looks high, the problem is not the budget: it is that one of the four numbers does not hold up.
Why does a budget that is too small not work?
Because campaigns need volume to optimize. Automated bidding systems learn from conversions, and with two or three a month they have no material to tell signal from coincidence.
The typical result of an insufficient budget is not “proportionally fewer sales”: it is zero learning. You spend the money and end up not knowing what was working, which is worse than never having started.
If the arithmetic gives a number you cannot sustain, the answer is not to spend less on everything: it is to narrow the scope — fewer keywords, one city, one service — so the same budget concentrates where volume is reachable.
How long before you judge it?
Three months minimum in B2B. The first month goes on the system learning and on fixing the obvious: irrelevant search terms, ads that do not match the page, conversions configured wrong.
The second month works on real data. The third is the first that says anything about profitability. Switching a campaign off in week three throws away the learning investment right before collecting on it.
What wastes the budget?
Four leaks account for most unproductive spend:
| Leak | Symptom | Fix |
|---|---|---|
| No conversions configured | It optimizes for clicks, not for business | Configure the event before switching the campaign on |
| Broad match with no control | You show up in searches that never buy | Review search terms and exclude weekly |
| A generic landing page | Good click, zero conversion | One page per promise the ad makes |
| Leads that never reach the CRM | Nobody knows which ones closed | Connect the form to the CRM and send the sale back to the campaign |
The fourth is the most expensive and the least visible: without returning the commercial result to the platform, you optimize for forms filled in rather than customers won, and those are not the same thing.
Google or Meta first?
Whichever holds the intent. Google captures whoever is already looking for a solution: high intent, high cost per click, short cycle. Meta interrupts someone who was not looking: lower cost, lower intent, longer cycle.
For a B2B service people search for by name — “CRM implementation,” “e-commerce development” — Google usually pays off sooner. For a category the buyer does not know exists, Meta educates better. Starting on both with the budget split down the middle is the most common way to learn nothing from either.
And if the campaign brings leads that do not buy?
Then the problem is rarely in the campaign. It is in the ad's promise, in how the form qualifies, or in the sales follow-up. The first thing we look at in that case is not the Ads account: it is what happens to the lead after the form is filled in.
A well-targeted campaign feeding a broken sales process produces exactly what the complaint describes: volume that does not convert, and a sterile argument between marketing and sales about whose fault it is.
So before recommending a budget increase, we review the whole path from campaign to sale: the missing money is almost never in the ad spend.