Paying for Ads on Search: A First-Time Approach

What a customer is worth, before opening an account

The first question is not which platform to use. It is what a customer is worth over the period you are prepared to wait for, and the answer belongs to finance rather than to marketing. Somebody has to look at average order or contract value, the gross margin on it, repeat behaviour if there is any, and how long a customer stays. Where the marketing team and the finance team disagree about any of those figures, that disagreement has to be settled first, because every later decision depends on it.

Three numbers are worth writing down at the start.

• Maximum cost per acquisition — what you can pay for a customer and still be in profit. Expressing it as a share of gross margin keeps a lower-margin business from assuming the same ceiling as a higher-margin one.
• Break-even cost per click — that figure divided by the number of clicks a converted customer typically needs. It is an estimate and it will be wrong, but it stops the account being judged on revenue per click against a target nobody agreed to.
• How long a conversion takes to close. A lead that becomes a customer after four months cannot be judged on a fortnight of data, whatever the dashboard says.

Put these on the first page of the account plan. Almost every argument the team will have later — raise the budget, change the bidding, try that keyword — resolves itself once there is a number at the top of the document.

Decide what the ads are responsible for

Paid search and organic search compete for the same queries, and the two are not additive. Running both on the same terms splits the work between channels, so the useful question is what the ads buy that organic will not deliver on its own.

There are several legitimate answers. Paid covers demand that will spike before content can be written, indexed and ranked. It buys presence on terms where a competitor’s organic position makes the paid position hard to take. It reaches segments the site does not address, such as a specific role inside a specific sector, or a particular competitor’s customers. And it gives you somewhere to test a message before committing six months of editorial work to it.

There is also an answer that is not legitimate: paying for traffic because somebody has decided that is what marketing is for. If nobody can name the job the ads are doing, the budget is a habit, and habits are expensive.

Where both run together, coordinate them rather than isolating them. Note the overlap in the plan so the same query is not being bid for by one team and worked on by another with different success measures. The channels also feed each other: an advert that brings people in and a page that converts them is a paid campaign resting on organic work. Our Google Ads management work starts by mapping that overlap rather than by opening the bidding screen.

Build an account you will still understand in six months

Account structure is the least glamorous decision and the one with the longest half-life. A structure mirroring how the business sells makes weekly optimisation possible. One mirroring how the account happened to be set up on day one makes it unreadable inside a quarter.

Four rules that have held up repeatedly.

• Campaigns divide by intent, not by keyword volume. Someone searching for a service they intend to buy is in a different market from someone researching a problem. Mixing them means you cannot read the results and cannot write the right landing page for either.
• Ad groups hold keywords that need the same ad and the same destination. If two keywords need different copy they are different ad groups. If they need different pages they are different campaigns.
• Keep geography and scheduling at campaign level unless you have a specific reason to move them. Everywhere-and-all-hours as a default on a local business account is a common and expensive oversight.
• Turn on conversion tracking before launch, not on the second day.

That last one needs checking rather than trusting. Phone calls from a mobile number are frequently not tracked. Form fills on a single-page site often fire twice, once on submission and once on the thank-you page. Imported offline conversions can duplicate everything if the matching is loose. Each of those errors inflates reported results, and inflated results are the reason budgets get raised on a campaign that is not delivering. Where the customer journey runs through a CRM, importing real outcomes back into the account is the task that matters most and the one most often deferred.

Screen the keywords yourself

Keyword tools are good at volume and bad at intent. They will report a term as worth bidding on because it converts adequately somewhere else entirely.

Read the terms before the bid column. Watch for the plural trap, where an agency bidding on a software product name ends up paying for a recruitment firm. Watch for support, free or tutorial modifiers, which rarely convert in a commercial campaign even when the clicks are cheap. Watch for the year qualifiers and the brand comparison terms, some of which are buyers and some of which are researchers.

Bidding on competitors’ brand terms is a real technique and often an expensive one. It only makes sense when you have something genuinely different to say. Running it by default because somebody else did it first is how budgets disappear.

For a first campaign, a short, hand-checked list of commercial-intent terms usually beats broad match with heavy negative management, and it teaches the team what customers actually type. Broad match has its place, usually once there is enough conversion history for the platform to learn from and a tight negative list to keep it honest. In week one there is no history, so there is nothing to learn from.

Where keyword research overlaps the organic plan, agree who owns which query. The same search engine optimisation work that lifts a term in organic should not be fighting the paid campaign for the same landing page.

The search terms report is where budget leaks

Once a campaign runs, the words that actually triggered your ads will never match your keyword list exactly. Some are irrelevant, some are surprisingly good, and some are questions that suggest an ad group nobody wrote. This report is reviewed weekly for the first few months and is usually the highest-value half hour a manager spends.

1. Sort by cost rather than by clicks. A cheap irrelevant term costs nothing; an expensive irrelevant term costs the campaign.
2. Add the obvious rubbish as negatives straight away, at a match level that will not catch a useful variation of the same phrase.
3. Promote the good surprises into real keywords and their own ad groups.
4. Look for a pattern. If the same irrelevant phrasing keeps appearing, the ad copy is promising something the landing page does not deliver, or the keyword list is too loose.

Two cautions. Negative lists grow into something that over-reaches if nobody owns them, and a term blocked at phrase level once stays blocked for months. And a query that looks irrelevant often is not: how much does this cost is frequently the highest-intent search in the account, whatever it looks like in the report.

Budget, bidding and the period where nothing works

Two mistakes do most of the damage in the first few months. Setting a daily budget far below what the market charges for a click produces a campaign that spends slowly and learns nothing. Spreading a fixed monthly figure by doubling the daily budget on quiet days is slightly better and still distorts delivery. Setting the budget to roughly what the market can absorb, accepting that it will be spent, is usually healthier — while watching spend against the acquisition ceiling agreed earlier.

Automated bidding needs a certain volume of conversions before it can do anything useful. Below that, manual or cost-per-click bidding gives more control and, more usefully, more information about what the market is charging. Switching to an automated strategy at week three because the campaign feels stuck is a common way to lose sight of the true cost of a click.

Ad scheduling deserves suspicion as well. Cutting an account to business hours saves nothing on cost per click, and it removes the data that tells you which hours convert. For a business with a long sales cycle that can quietly remove conversions you did not know you had.

Whatever the settings, agree who is allowed to change them. Budget changes made by somebody who cannot see the acquisition ceiling are the most common way an account overspends in its first six months, and SmartEdge IT Solutions records that agreement with the account notes rather than relying on memory.

Measurement that survives contact with the sales team

The distance between what the platform reports and what the business receives is usually the widest gap in a paid search programme, and it is worth closing deliberately rather than arguing about later.

Begin by agreeing the definitions in writing: what counts as a conversion, which of them are worth bidding towards, and which are noise. A form fill and a phone call may both matter; a page view will not. Then verify the tracking matches those definitions on a real device rather than in the platform’s own preview.

Where the outcome is only known later — a lead that becomes a customer, a trial that converts — feed it back. Export outcomes from the CRM and import them as offline conversions carrying the campaign and keyword recorded at the original session. That takes API work between the CRM and the advertising account, and it changes bidding decisions more than anything else on this list, because the platform can finally optimise towards the thing you actually sell rather than the thing somebody recorded as a goal.

Report cost per accepted lead and cost per customer next to cost per click and cost per conversion, in the same place — often a simple reporting dashboard — so nobody gets to choose which number to believe. Attribute the first click honestly: a typical search journey touches several queries before a form submission, so some assumption is unavoidable, and stating it beats pretending otherwise.

A workable first quarter

Months one and two are for learning, and the goal is information rather than return. One campaign per intent, one ad group per distinct need, hand-checked keywords, tracking verified on a real device, a weekly search terms review, and a written note of what each test taught you.

Month three is when decisions get made. By then you should know your actual cost per lead, which segments convert, and the queries you did not anticipate. From there the moves are straightforward: prune the campaigns that consistently produce leads nobody wants, scale the ones producing leads that close, and test a new landing page against the existing one rather than adding keywords to an unchanged page.

Keep the account notes current throughout. A paid search account is a record of hypotheses tested, and in a year’s time that record is worth more than any single campaign. It also matters for whoever inherits the account later, which is a more common event than teams like to assume. SmartEdge IT Solutions hands over that record with the account, because an undocumented account is an account nobody can safely change.