Where your PPC budget goes

An owner who has never run paid search asks a reasonable question and gets an unreasonable answer. How much does advertising cost? The reply is that it depends. It does depend, but the components are known, and once they are laid out the figure stops being a mystery.

What an advertising budget is made of

The budget is made up of five items, and they behave very differently.

• The money paid to the platform. Spent daily, usually on clicks, though in some formats it can go on impressions instead.
• The cost of running the campaigns. Someone builds, watches and adjusts them.
• Preparing the destination page. Where the traffic lands, if it does not exist yet.
• Setting up measurement. Without it the rest cannot be judged.
• Creative work. Text, images, product feeds where they apply.

The first two are the ongoing spend. The last three are mostly one-off, and skipping them is the most common way to waste the first two.

The proportion between platform spend and management fee shifts with scale. On a small budget the fee is a significant share, which is why very small budgets rarely justify an agency. On a larger one the fee becomes a minor line and the platform spend dominates.

The cost of a click and what it depends on

The platform does not have a price list. Each click is priced by an auction that runs at the moment somebody searches.

What you pay depends on how many advertisers want the same query, how relevant your advertisement is to it, and how good the page behind the advertisement is. That last part surprises people. A slow page or one that does not answer the query raises what you pay per click for the same position, because the platform judges the whole experience rather than just the bid.

This is the practical reason to fix speed before raising the budget, because a slow page is paid for twice, once in the auction and once in the visitors who leave before it renders. Where that slowness usually comes from is covered in our article on why a website loads slowly.

Time also matters. The same query is priced differently on a Monday morning and a Saturday night, and in some trades the difference within a single day is large enough to justify running the campaign only during certain hours.

Why a click costs so much more in some trades

The spread between trades is wider than any newcomer expects, and it has a straightforward logic.

The rule underneath all of it is that a click costs roughly what the trade can afford to pay. Where one customer brings in several thousand, advertisers bid accordingly. Where the average order is small, the auction settles low because nobody can pay more.

That is why comparing your click price to a figure someone quoted for a different trade tells you nothing. The only comparison that means anything is against your own numbers, which is the arithmetic in a later section.

Where a click goes and why not every one becomes an inquiry

A click buys a visit, and the distance between a visit and an inquiry is where most budgets are lost.

Out of a hundred people who click, a proportion leave within seconds because the page is not what they expected. Another group reads and leaves because the offer does not suit them. Another was researching rather than buying. What remains is the group that makes contact, and in most trades that group is small enough to be measured in single-digit percentages.

None of that is a fault by itself. It becomes a fault when the loss happens for reasons you control.

• The page does not match the advertisement. Different service, different promise.
• The page is slow on a phone. Most paid traffic is mobile.
• The offer is hidden below the fold. You have to scroll to find the point.
• There is no clear next step. Nothing obvious to press.
• The form is long or broken. The most expensive fault of all.

A broken form is common and invisible. The campaign runs, the money is spent, the reports show clicks, and the inquiries were never delivered. Testing the form by actually sending one from a phone takes two minutes and should be done before every campaign and after every change to the site.

What the page needs to turn a paid click into an inquiry rather than waste it is described in our article on Google Ads landing page requirements.

How much a test really needs

This is the question owners most want a number for, and the honest answer is a method rather than a figure.

A test is enough when it has produced sufficient inquiries for the decision you are making, not when it has run for a certain number of days. Seeing whether inquiries arrive at all takes less data than putting a figure on what one inquiry costs, so the volume of data is worked out for the campaign in front of you. What you can afford to pay for one inquiry is set by your margin, and on a handful of inquiries any conclusion is down to chance.

There is a second condition that gets skipped. The test is only valid if the measurement was correct from the first day, and setting that up after the campaign has been running for two weeks means those two weeks cannot be included. Getting the conversion tracking right before launch is one of the small tasks that sit inside Google Ads management, and it takes an afternoon rather than a project.

So the calculation runs backwards. Estimate what a click costs in your trade, gauge what share of visitors will make contact, and divide the click price by that share to find what one inquiry costs. Multiply that by the number of inquiries you need for a verdict, and you have the test budget.

The uncomfortable part is what happens when that figure comes out large. It usually means either the trade is expensive and the test genuinely costs what it costs, or the expected conversion rate is too low, which is a page problem rather than an advertising one. Splitting a budget that is too small across three campaign types produces three inconclusive tests instead of one useful one.

Search, shopping and remarketing

The three main formats behave differently and are worth separating in the budget.

Search ads appear in response to a query and reach people who are actively looking. They are the most expensive per click and usually the most productive, because the intent is already there.

Shopping advertisements show a product with a price and a picture, and they apply to retail. The person sees the price and the store name before clicking, so the visit is a more considered one. They require a properly maintained product feed. If the price or availability in the feed does not match the page, Google may disapprove the product, and repeated mismatches can lead to the Merchant Center account being suspended.

Remarketing reaches people who have already visited. It is the cheapest of the three per click, though some of its placements on other sites are paid for by impression, not by click. It is also the easiest to overdo. Shown moderately it recovers people who were interrupted; shown constantly it irritates them and burns budget on people who had already decided against you.

The usual mistake is starting with all three at once on a small budget. Starting with the one that matches the intent closest to a purchase and expanding once it works is slower and considerably cheaper.

Display advertising sits apart from those three. It is banner space on other sites, and the audience is the coldest of all because nobody was searching for anything. This is the easiest place of all to spend a budget with nothing to show for it, so it is added later and carefully, once search and remarketing are working.

Negative keywords and why a budget melts away without them

A campaign without negative keywords pays for clicks from people who were never going to become customers.

The platform matches your keywords loosely by default, which means an advertisement for paid services appears for queries containing words like free, do it yourself, second-hand, salary, or the name of a competitor’s product you do not sell. Every one of those is a paid click with no possible outcome.

Building the negative list is not a one-off task. The list of what people actually searched before clicking is available in the account, and reading it weekly in the first month, then monthly, is where a large share of the waste is found and removed.

In a campaign that has never had its search terms reviewed, it is common to find that a substantial part of the spend went on queries the owner would have excluded instantly had anyone shown them the list.

Costs noticed only after launch

Several lines rarely appear in the initial estimate and reliably appear in the invoice later.

• Handling the inquiries. Somebody must answer them promptly.
• Page changes. The first data usually shows something needs fixing.
• Extra landing pages. Different services need different destinations, and what one of them costs is worked through in our article on the cost of a landing page.
• Analytics work. Getting reliable numbers takes setup and maintenance.
• Seasonal bid increases. Competition rises before peak periods.

The first item is the one that undoes campaigns. Inquiries that sit unanswered for a day convert far worse than those answered within the hour, and in a competitive trade the customer often writes to several companies at once. Paying for clicks and then answering slowly is paying for someone else’s customer.

Ongoing changes to the page and its measurement are part of what runs under project support and development, and what that covers month to month is described in our article on what website support includes. Advertising and site maintenance are difficult to separate in practice even when they are separate lines on paper.

How to calculate cost per inquiry rather than cost per click

The click price is the number everyone watches and the least useful one.

The figure that matters is what one inquiry costs, which is the spend divided by the number of inquiries. Below that sits the figure that matters even more, which is what one customer costs, because not every inquiry becomes a sale.

• Cost per inquiry. Total spend divided by inquiries received.
• Share of inquiries that become sales. From your own records, not an estimate.
• Cost per customer. The two figures above, combined.
• Average order value. What a customer is actually worth.
• Repeat purchases. Whether the first sale is the only one.

With those five numbers the question of whether advertising works answers itself. Without them, the discussion stays at the level of impressions and clicks, which is a discussion about activity rather than about money.

The comparison people forget is against other channels. With advertising you pay for every click, while organic search has no per-click cost but needs ongoing work. That does not make advertising the worse option, because search results take months and advertising works this week. It does mean the two should be judged on different horizons.

When advertising is better paused

Turning campaigns off is a legitimate decision and it is made too rarely.

The clearest case is when the site cannot serve the traffic. If the page is broken, the form does not deliver or the product is out of stock, every click is money burned, and pausing until it is fixed loses nothing.

The second case is a cost per inquiry that has climbed above what a customer is worth, and has stayed there after adjustments. Continuing in that state is paying to lose money more quickly.

The third is capacity. A workshop booked out for six weeks does not need more inquiries this month, and the budget is better held until it does.

What does not justify pausing is a bad week. Weekly figures fluctuate, and reacting to every dip by changing bids produces a campaign that never stabilizes long enough to be judged.

Conclusions

An advertising budget is not a single number. It is platform spend, management, a page worth landing on and measurement that works, and removing any of the last three makes the first one less effective rather than cheaper.

The click price is set by the auction, your trade and the quality of your page, and nobody can negotiate it. The cost of a customer is far more in your hands, because much of it depends on how convincing the page is and how quickly you answer.

If you are running campaigns and cannot say what one inquiry costs you, send us the site address and access to the advertising account. We will check that the conversions are actually being recorded, look through the search terms the budget went on, and come back with what is worth fixing on the page before any bid is touched.