There is a keyword in a real Google Ads account with a Quality Score of 3 out of 10. By every piece of advice on the internet, that keyword is a problem. It is dragging down the account, it is costing more than it should, and it needs fixing or pausing.
That keyword produced 81 phone calls. More than anything else in the account.
The advertiser running it had a choice most owners face every week without realizing it: believe the number Google put on the screen, or believe the phone. He believed the phone, left the keyword alone, and kept the best performer he had.
That is the whole problem with Google Ads reporting in one story. The platform gives you a column for almost everything, and the columns owners stare at hardest are usually the ones that cannot actually decide anything.
The Column That Grades You Isn't the Column That Pays You
Every number in your account is doing one of two jobs.
Some numbers decide. They tell you whether the money you spent came back. Pausing a keyword, raising a budget, killing a campaign, doubling down: those calls get made on these and nowhere else.
The rest explain. Quality Score, click-through rate, cost per click, impression share. They tell you what is happening inside the machine so you know which part to go fix. They are genuinely useful. They just cannot tell you whether you are making money, which means they can never tell you what to switch off.
Deciding numbers and explaining numbers. Once you have that split in your head, most of the confusion in an ad account dissolves, and so does the temptation to act on the wrong thing.
The 3 out of 10 keyword makes sense immediately through that lens. Quality Score is an explaining number. It was telling the advertiser something real, that Google considered the match between the search, the ad, and the page less than ideal. But the deciding number, cost per phone call, was excellent. When an explaining number looks ugly and the deciding number looks good, the explaining number loses. Every time.
There is also an order to the work. Read the account first, form a view, then act. Most owners do the opposite: they scroll, find something that looks wrong, and start changing settings before they understand what they are looking at.
Four Columns Decide. Everything Else Is Commentary.
If you generate leads, which is most local and service businesses, the four are amount spent, conversions, cost per conversion, and one level deeper than Google can see, cost per booked job. If you sell products, swap the last two for cost per purchase and conversion value against spend.
Put those to the left where you cannot avoid them, and sort by amount spent, highest first. Not by cost per conversion. Sort by where the money actually went, because a keyword with a beautiful cost per lead and eleven dollars of spend is a rounding error, and sorting by cost per conversion will float it to the top of your screen every week pretending to be your best performer.
All of this rests on one thing: conversion tracking has to be working. Without it none of these four columns exist, and you are left making decisions on clicks and impressions, which is the same as making them at random. If you are not tracking conversions, stop reading about metrics and go fix that first. It is the single highest-value hour you will spend on the account.
One more thing about the fourth column. Google can tell you what a lead cost. It cannot tell you whether that lead answered the phone, showed up, or bought anything. Those two numbers diverge more than most owners expect, and we come back to it at the end.
Quality Score Is a Diagnosis, Not a Grade
Quality Score is a 1 to 10 rating on a keyword, built from three things Google is judging: how likely people are to click your ad for that search, how well your ad matches the intent behind it, and whether your landing page delivers what the ad promised. We pulled it apart properly in A Bigger Bid Won't Win You the Top Spot. Here's What Does.
It is not in your columns by default. You add it, which tells you something about how central Google thinks it is to your day-to-day decisions.
Here is how to hold it. A low Quality Score is a signal about relevance, and relevance is worth improving, because a better score generally means Google charges you less to compete for the same searches. The usual fix is not mysterious: pull the mismatched keywords into their own ad group, write ads that use the actual words people are searching, and point them at a page about that specific thing rather than a general services page. If you sell both roof repair and gutter cleaning through one ad group pointed at one page, you should expect a mediocre score, because you have earned one.
But a low score on a keyword that is producing cheap leads is not a problem to solve. It is a curiosity. Improving it may make your leads slightly cheaper. Pausing the keyword because the number offends you costs you the leads.
Where Quality Score earns its keep is on keywords with no conversions yet. Early in a campaign, before you have enough conversion data to judge anything, it is a reasonable proxy for whether the setup is sane. That is its real job: an early warning while you wait for the numbers that actually decide.
A Low Click-Through Rate Is a Message, Not a Verdict
Click-through rate is the share of people who clicked after seeing your ad. On search campaigns it runs much higher than people expect, because someone typing "emergency plumber near me" is already looking for you. Practitioners get uneasy below about 2% on a search campaign and consider anything approaching 10% strong, but those are rough guides from other people's accounts, not benchmarks you should hold yourself to.
What matters is what a low one is telling you, and it is almost always one of two things.
Your ad copy is not earning the click. The headline is generic, it does not name what the person searched for, or it reads like every other ad on the page. This is the more common cause and the easier fix.
Your keywords are pulling in the wrong searches. If you are matching broadly, plenty of people are seeing your ad for things that are only loosely related to what you do. They scroll past, correctly, and your click-through rate reflects it. That is not an ad problem, it is a targeting problem, and tightening match types or adding negatives fixes it. We went through that in You're Paying Google for Searches You'd Never Choose.
Same symptom, opposite fixes. Which is why a click-through rate on its own is a message rather than an instruction.
Cost Per Click Is a Price, Not a Score
Plenty of advertisers treat a falling cost per click as progress. Sometimes it is. Often it just means you drifted toward cheaper, less valuable searches.
Cost per click is set by what other businesses will pay to reach the same person for the same search. Some keywords in your industry are expensive because they are the ones that turn into customers, and no amount of optimization is going to make "commercial roofing contractor" cost the same as "roof cost calculator." Improving your ads and your relevance genuinely helps, and it is worth doing. It will not repeal the auction.
The lever most owners never pull is the keyword list itself. If a slice of your keywords sits in a much more expensive corner of your market, the question is not always how to bid smarter on them. Sometimes it is whether there are less contested searches that convert nearly as well, which you can only answer by looking at conversions per keyword rather than cost per click.
Impression Share Tells You How Much Room Is Left
Search impression share is how often your ads showed out of the times they were eligible to. Like Quality Score, you have to add it.
It is the closest thing Google gives you to a headroom gauge, and the read is refreshingly simple.
Low impression share, say 20 or 30%, means most of the demand for your keywords is going to someone else. If your cost per lead is good, that is not bad news, it is the best news in your account: there is more of this available and you are not taking it.
High impression share, up in the 80s or 90s, means you are already capturing nearly everything these keywords have to offer. More budget will not buy much more, because there is not much more to buy. Growth from here has to come from somewhere else: new keywords, new ad groups, new campaigns, a wider service area.
That distinction matters because it answers a question owners ask constantly. "Can I scale this?" Impression share answers it in about four seconds, and it is a far better answer than guessing.
Lost to Budget and Lost to Rank Point in Opposite Directions
This is the pair almost nobody uses, and it is the most decision-useful thing in the whole reporting suite.
When you do not show, Google tells you why, split across two columns. Search lost impression share (budget) is the share of eligible impressions you missed because your budget ran out. Search lost impression share (rank) is the share you missed because your ad was not good enough to win the auction.
They sound similar. They point at completely opposite fixes.
Lost to budget is a money answer. You were eligible, you could have competed, and you ran out of road. If the campaign is profitable, this is a green light. Raise the budget, watch what happens to your cost per conversion, and raise it again. The number tells you roughly how much is sitting on the table: losing 20% to budget means there is meaningfully more volume available at broadly similar economics, until you start eating into it.
Lost to rank is an auction answer, and raising your budget does nothing at all. You were eligible, you competed, and you lost, because Ad Rank decided other ads deserved the slot. The distinction that matters here is between budget and bid. Budget is how much the campaign may spend in a day, and it is irrelevant to an auction you already entered and lost. Bid is what you are willing to pay for the click, and that genuinely is part of Ad Rank, so bidding higher can win some of those impressions back. It just wins them back at a higher price forever.
The cheaper fix is relevance, the other half of Ad Rank: split the ad group so each one covers a tighter set of searches, put the words people actually type into the ad, and send the click to a page about that specific service instead of a general one. That buys you the same impressions without paying more for every click you were already winning. When practitioners see this number up in the 80s or 90s, the diagnosis is almost always the same, that the campaign is too broad and the keyword, the ad, and the page are not saying the same thing.
Both near zero is its own answer. You are winning nearly everything you are eligible for. Nothing left to squeeze. Growth means expanding what you are eligible for, which means new keywords and new ad groups, not new money on the old ones.
Your Conversions Are Filed Under the Day of the Click, Not the Day of the Sale
This one causes more unnecessary panic than any other number in the account, and almost nobody explains it.
When someone clicks your ad on Monday and becomes a customer on Friday, Google files that conversion under Monday, the day of the click, not Friday when it actually happened. The conversion travels back in time to sit next to the click that caused it.
Which is the right way to measure an ad. It is a confusing way to read a recent week.
Look at the last seven days on a business where people take a while to decide, a roofer, a law firm, a dentist doing implants, and it will look worse than reality every single time, because the clicks are recorded and the conversions they will produce have not arrived yet. Owners see that, conclude the account fell off a cliff, and start changing things that were working fine.
Two habits fix it. First, judge recent performance on longer windows and expect the last week to fill in. Second, if you want to see what actually happened in a period rather than what a period's clicks eventually produced, Google offers conversion columns counted by conversion time instead of click time. Adding one of those next to the standard column shows you both views, and the gap between them is a decent measure of how long your customers take to make up their minds.
If the two columns are nearly identical, your customers decide fast. If they diverge a lot, they do not, and you should stop judging any window shorter than your typical decision cycle.
Read Them in Pairs
One number on its own rarely tells you what to do. Two together usually do, because the same symptom has several possible causes and the second number sorts them out.
In plain English:
Good click-through rate, poor conversion rate. Your ad is doing its job and the page is not. People arrived interested and left. That is a landing page or offer problem, and no amount of keyword work fixes it. Check that the page delivers what the ad promised, because a mismatch there kills more campaigns than bad targeting ever has.
Poor click-through rate, good conversion rate on the few who come. You are reaching the wrong people but converting the right ones well. Tighten the targeting rather than rewriting the offer.
High cost per conversion, everything else normal. Now you are usually looking at the offer itself, or the price, or the honest possibility that the market wants this less than you hoped. Harder to hear, and much cheaper to learn now than in six months.
Low impression share plus lost to budget, and good economics. Nothing is broken. You are just under-funding a winner.
Low impression share plus lost to rank. A bigger budget changes nothing here. Fix relevance first, and only then consider bidding more.
Low Quality Score, good cost per conversion. Leave it alone. This is the 3 out of 10 keyword with 81 phone calls.
The Only Number Your Bank Cares About
There is one level below all of this, and it is where the businesses that grow on Google Ads separate from the ones that keep score on Google Ads.
Cost per lead is not cost per customer. Thirty leads at forty dollars looks better than twelve at ninety until you find out the forty dollar leads were price shoppers who never called back and the ninety dollar leads booked. Google cannot see any of that. It sees a form submitted and a phone ringing, and its job ends there.
So take it one step past the platform. Track which leads turned into quotes and which quotes turned into jobs, even if that is a spreadsheet you fill in on a Friday afternoon. Then divide your spend by booked jobs. That number, cost per booked job, is worth more than every column in this article put together, because it is the only one measured in the same units as your bank account. It is also the number that tells you which keywords deserve more money, which is a very different list from the one with the best Quality Scores.
None of this requires you to become an analyst. Four columns to decide with, a handful more to explain with, read in pairs, checked against your own history rather than someone else's benchmark. Fifteen minutes a week, and it is the difference between running an account and watching one.
If you would rather someone else did that fifteen minutes, with the follow-through all the way down to booked jobs, that is what we do all day. Have a look at how we run Google Ads, or keep the tab open and work through your own columns. Either way, sort by amount spent first, and be slow to punish a keyword that is bringing in the phone calls.
