Google Search Ads management
A Search campaign only works if five separate things agree with each other: the query you bid on, the intent behind it, what the ad says, what the page shows, and what you count as a conversion. Break the chain anywhere and the account still spends. It just stops producing customers.
I manage that whole path rather than the campaign settings in isolation. Most of the accounts I take over are competently built and quietly leaking, usually somewhere between the query and the page.
Google Search Ads place a text ad in the results when someone searches for something you have chosen to bid on. The whole model rests on one fact: the person typed the words themselves. You are not persuading someone to want a plumber. You are trying to be the plumber they pick.
That changes what the work is. On channels where you interrupt people, the message has to create the interest. On Search, the interest already exists and the job is to be relevant to it, at a price the business can carry. Relevance is not a soft quality here. It is the thing that decides whether a click was worth buying.
It also puts a ceiling on the channel. Only so many people search for what you sell in a given month, and no amount of budget creates more of them. When a Search account plateaus, that ceiling is often the reason, and the honest answer is that growth has to come from somewhere else rather than from raising bids.
Search rewards precision in a way other campaign types do not. A single badly chosen keyword can absorb a month of budget on queries that were never going to buy, and the account will report those clicks as normal activity.
Capture demand, or create it
Neither is better. They answer different questions, and a business with no existing search demand for its category will not fix that with a Search campaign.
I work the path in order, from the query a real person typed to the money that lands in the business. Seven stages, and a fault at any one of them shows up as the same complaint: Google Ads is expensive and it is not working.
Two people can type nearly the same words and want opposite things. Someone searching "CRM software pricing" is shortlisting. Someone searching "what is a CRM" is reading. The keyword tool shows both, ranks the second one higher for volume, and says nothing about which one buys. Sorting that before building anything is the difference between a lean account and an expensive one.
The first intent map is a set of assumptions. Once real conversions arrive, some of those assumptions turn out to be wrong, and usually in useful ways: a term written off as research traffic converts, and a term that looked commercial brings tyre kickers. Feeding that back is the part that compounds.
Waste in a Search account is rarely one big mistake. It is usually four or five small ones that each look defensible on their own. Here is what I check first, and what each symptom tends to point at.
| What you see | What it usually means | What I inspect first |
|---|---|---|
| Spend is up, enquiries are flat | The account is matching queries that read as relevant but carry no buying intent | Search terms report over 60 to 90 days, sorted by cost with zero conversions |
| Good click volume, poor conversion rate | The ad promises something the landing page does not confirm quickly enough | Each ad group against the page it points to, on a phone, at normal connection speed |
| Sales says the leads are weak | The account is optimising towards form fills because nothing tells it which ones closed | Whether CRM outcomes are fed back as conversions, and how leads are valued |
| Cost per conversion jumped without a settings change | Auction pressure, seasonality, or a broadened match picking up new queries | Search terms added in the period, average CPC trend, impression share lost to rank |
| Bidding target is set but never met | The target was chosen by convention rather than derived from margin, or data volume is thin | How the target was arrived at, and conversion volume in the campaign per week |
| One campaign consumes the whole budget | Budget sits at campaign level while several campaigns chase the same terms | Overlap between keyword sets, and budget-limited status by campaign |
| Negative keyword list has not grown in months | Nobody owns search term review, so waste accumulates quietly | Date of the last negative added, and cost of non-converting terms since then |
| Reported conversions look healthy, revenue does not | A conversion action counts something short of a sale, or counts it more than once | Conversion action definitions, counting settings, and a period reconciled against orders |
The pattern across most of these is that the account is doing exactly what it was told to do. Nothing is broken in the sense of throwing an error. The instructions were just wrong, and the platform followed them efficiently.
If several rows above sound familiar, the wider set of causes is worth reading first. I have written up why Google Ads may not be working as a diagnostic walkthrough, and a full Google Ads audit puts numbers against it for your specific account.
Six areas of ongoing work. They map onto the stages in the path above, which is deliberate: the management is the methodology applied week after week rather than a separate checklist.
Search intent and keyword architecture
Keywords get grouped by the intent behind them, not by topic. A comparison term and a ready-to-buy term belong in different places because they need different ad copy, different pages, and often different acceptable costs. Match types set how wide the net goes for each group, and that choice gets revisited as search term data arrives rather than fixed at launch.
Search terms and negative keyword control
The search terms report is the only place you can see what you actually paid for, as opposed to what you asked for. I review it on a schedule, add negatives at the right level so one campaign's exclusions do not silently block another, and keep a running note of which query patterns never convert. This is unglamorous and it is where a good chunk of recovered budget comes from.
Responsive Search Ads and message relevance
Google builds each ad by combining headlines and descriptions you supply, so the assets need to be varied enough to give it real options and consistent enough that any combination still makes sense. Pinning solves specific problems, such as a legally required phrase, and creates others when overused because it removes the combinations the system would have tested. Ad Strength is a guide to asset variety, not a score to chase for its own sake.
Landing-page alignment
Every ad group should point at a page that confirms its promise in the first screen. Where the right page does not exist, I say so rather than routing the traffic somewhere generic and hoping. I do not always control the site, so this often arrives as a specific recommendation with the reason attached, which your developer or content team can act on without a meeting.
Conversion tracking and value signals
Conversion actions get checked before anything else, because automated bidding spends against them. That means confirming each action fires once at the right moment, carries a value where a value exists, and reconciles with your own order or enquiry records for a fixed period. Setup usually runs through Google Tag Manager alongside Google Analytics 4, with Enhanced Conversions where it applies. For lead generation, this is also where CRM outcomes come back in so the account can tell a good lead from a fast one.
Bidding and budget decisions
Smart Bidding options such as Target CPA, Target ROAS, Maximize Conversions and Maximize Conversion Value each need different things from an account. The choice depends on how much conversion data the campaign actually generates, how stable that data is, and whether you have a defensible target derived from margin. Where the data is thin, more manual control is usually the better answer until it thickens. Budget follows the same logic: it goes where the evidence supports it, not evenly.
Where an account needs work across several campaign types at once rather than Search alone, that sits under broader Google Ads management.
Four tiers, read from the top down. Tier one is what the business feels. Tier four is a set of diagnostics that help explain tiers one to three, and that is all they are. Reports that lead with tier four are describing activity rather than results.
1
What the account is accountable for. If these do not move, nothing below matters.
2
The bridge between the account and the business. Useful only when the conversion actions behind them are trustworthy.
3
Where the money went and why. These explain movement in tier two, and the search terms report earns its place at the top of this list.
4
Keyword-level indicators, useful for finding a mismatch between a keyword, its ads and its page.
Google describes Quality Score as a diagnostic on a one to ten scale, meant to show how your keyword compares with others. It is not what enters the auction. Ad Rank is, and it takes in your bid, ad quality at auction time, the context of the search, and expected impact from assets and formats. Quality Score is a rear-view estimate built from historical data.
I use it the way a mechanic uses a warning light. A keyword sitting at three with weak ad relevance is telling me its ads or its page have drifted from what it is matching. That is worth investigating. Raising the number is not a goal in itself, and any promise that improving Quality Score will cut your costs by a stated percentage is being made up.
The mechanics stay the same. What changes is what counts as a conversion, how long you wait before you can trust a number, and how much a wasted click costs you.
Ecommerce
Search handles the queries product data does not express well: category terms, comparisons, "best" and "vs" searches, and brand defence. Purchases are measurable within days, which makes decisions faster than in any other model. The catch is that a good account-level return can hide items sold below their real cost once goods, shipping and returns are counted. Feed-driven campaign types and product-level economics belong to Google Ads for ecommerce rather than here.
B2B and lead generation
A single click can be worth a great deal, so a high cost per click is often perfectly rational. The difficulty is the lag and the proxy: a deal may close months later, and a form fill is a weak stand-in for a customer. Without sales outcomes coming back from the CRM, the account learns to buy whoever fills in forms most readily, which is usually not the buyer you wanted.
Local services
Intent is often urgent and the market is finite. Someone searching for an emergency electrician at eleven at night is not comparing five quotes. Calls and messages matter more than form submissions here, and they are the conversions most often left untracked. Geographic settings and location-aware copy do more work than clever structure.
SaaS and technology
Most demand arrives as a described problem rather than a product name, so the query set skews towards symptoms and alternatives. A free trial is a weak conversion to optimise against because trials and paying customers behave differently. The target that actually helps sits at activation or first payment, which means measurement has to reach into the product, not stop at the signup form.
Professional services
Low volume, high value, and a long trust-building path. Small numbers make it easy to over-react: two quiet weeks look like a problem and are usually noise. Judgement here is mostly about patience and about accepting a cost per enquiry that would be alarming in a higher-volume business.
Each campaign type is good at one job. Problems start when a business expects one of them to do another one's work, then concludes the platform does not suit them.
| Campaign type | Primary role | Demand signal | Best fit | Important limitation |
|---|---|---|---|---|
| Search | Answer an active query with a relevant offer | Explicit: someone typed it | Products or services people already look for by name or problem | Volume is capped by how many people search |
| Shopping | Show product, price and image against product searches | Explicit, drawn from product data | Retailers with a catalogue and a clean feed | Only as good as the product feed behind it |
| Performance Max | Pursue a conversion goal across Google inventory | Mixed, decided by the system | Accounts with reliable conversion data and steady volume | Limited manual control and coarser reporting |
| Display | Place visual ads across sites and apps | Implied, from audience signals | Re-engagement and broad reach at low cost per impression | Attention is incidental, so quality varies widely |
| YouTube | Hold attention with video | Implied, from viewing behaviour | Brands with video that can earn a few seconds | Rarely converts on last click, so it gets misjudged |
| Demand Gen | Create interest before anyone searches | None yet, it is being generated | Products people do not know to look for | Slower to read, harder to attribute cleanly |
Search and Performance Max cause the most confusion because both can produce purchases and leads. The practical difference is control. In Search you decide which queries you compete for and you can read exactly what you paid for. Performance Max pursues a goal across Google inventory with far less manual say and coarser reporting. Which suits an account depends on how reliable its conversion data is and how much volume it produces, not on which is newer.
Search buys visibility immediately, which is its strength and its risk. It will expose an unresolved problem faster and more expensively than any other channel, so it is worth checking these before spending.
On that last point, I will not promise a cost per acquisition, a return on ad spend or a timeline. Those depend on your pricing, your competitors, your site and the auction, and I control none of them. What I will commit to is a defensible order of work, reporting you can check against your own records, and telling you early if I think the budget belongs elsewhere.
Four phases. The first one is not optional, because everything measured afterwards is read through it.
Phase one
Conversion actions verified against your own records, values checked, duplicates found. Then a written baseline of where the account stands, so later change can be argued about with evidence rather than impressions.
Phase two
Keywords regrouped by intent, negatives applied, ad copy rewritten where it does not match the query, landing-page gaps flagged. The highest-cost problems go first, not the quickest ones.
Phase three
Search term review on a schedule, asset testing, bidding adjusted as data accumulates. Changes made one at a time where possible, so results stay attributable to a decision.
Phase four
Reporting against the four tiers above, with a view on whether more budget is justified. Efficiency usually softens as spend rises, so the question is where the return crosses what your margin can carry.
Three situations that come up often, written as what was wrong, what it turned out to be, and what changed. I have left figures out rather than quote numbers I cannot show you the workings for.
Send me the account and a short note on what you sell. I will tell you where the money is going and whether the numbers you are reading can be trusted. If Search is not where your budget belongs, I would rather say so than build you a campaign.