A construction company's office manager types "construction crm" into Google. Two software ads appear. The first says "The CRM Small Businesses Love" and opens a home page about sales pipelines, marketing email and integrations. The second says "CRM Built for Contractors" and opens a page about bids, job sites and subcontractors.
Both companies may sell the same product. Only one of them answered the question. Google can tell the difference, and it charges accordingly.
This chapter covers how Google ranks ads and what that means for how a software account is built. In brief: money is one lever, and the cheaper lever is making the search, the ad and the page agree. Our evidence is the dozen project management, CRM, payroll and scheduling companies from earlier chapters, all advertising on US searches.
Ad Rank Weighs Your Ad and Your Page Alongside Your Bid. The Highest Bid Can Still Lose the Top Spot.
A search for "construction crm" sets off an auction in the time it takes the page to load, and every software ad in it receives an Ad Rank. Ad Rank decides whether the ad shows at all and where. The ingredients, as Google lists them:
- The bid, the ceiling you set on a single click.
- Ad and page quality, scored fresh on every search for that particular searcher.
- Ad Rank thresholds, the minimum an ad has to reach to show.
- The competition, since an ad ranked only slightly above its neighbors holds its spot less firmly than one far ahead.
- The context of the search: the person's location, device, the words they typed, the time.
- What your extras are expected to add, like sitelinks or callouts under the ad.
Notice where the bid sits on that list: first, and alone. Of the rest, your ad, your page and your extras are yours to improve; the competition and the searcher's situation are not. Because quality is part of the calculation, an advertiser with a better ad and a better page can outrank one with a bigger bid.
Google doesn't publish a formula for any of this, and any article that hands you one is guessing. The one promise Google does put in writing is about price: when your ad quality goes up, your cost per click often comes down.
Quality Score Is a Report Card, Not an Entry Fee.
Open the keywords tab in a software account and each row carries a number between 1 and 10, its Quality Score. Many advertisers treat that score as the thing the auction runs on. It isn't. Google is explicit that the score itself never enters the auction. It's a readout, built to show how your ad and page stack up against competing advertisers on that keyword.
It's still worth reading, because its three parts point to where the trouble is:
- Expected click-through rate. Google's estimate of whether people searching that keyword will click your ad.
- Ad relevance. Whether your ad matches what the searcher is looking for.
- Landing page experience. Whether the page behind the ad delivers on it for the person who clicked.
Expected click-through rate trips up a lot of software advertisers, because it sounds like something a higher bid would fix. It isn't. Google strips position and ad extras out of the measurement on purpose, so the score reflects the ad itself. Pushing your ad higher on the page doesn't raise it, and neither does adding sitelinks. What raises it is an ad that a person searching that exact keyword wants to click: their words in the headline, the answer they came for, a reason to pick you over the four other software ads on the page. For a buyer typing "crm for accountants," an ad that says "A CRM Made for Accountants" is the click; "All-in-One CRM Platform" is the scroll.
A low score on any of the three is a hint, not a penalty. The fix for each is different. The next sections start with ad relevance, then move to the landing page.
An Ad Group Built Around "Construction CRM" Beats One Built Around "CRM" on Ad Relevance.
Of the three, ad relevance is the easiest to improve, and the lever is how keywords are grouped.
A software account built carelessly has one ad group called something like "CRM" holding forty-odd keywords: "crm software," "crm for small business," "construction crm," "crm for accountants," "sales pipeline tool." One ad has to answer all forty, so it ends up saying something general enough to fit every one of them and specific to none.
Split it instead: each ad group holds the handful of keywords, five or six, that ask the same question, and its ad answers only that question. Some of those intents are bigger than you'd expect. Keyword Planner's US numbers put "construction crm" at about 880 searches a month, "crm for agencies" about 1,000, "call center scheduling software" about 390, and "church payroll software" about 110. Next to "crm software" at 165,000 they look small. But every one of those searchers has told you exactly who they are.
When the search, the ad's headline and the page all carry the same words, Google has an easy time judging the ad relevant, and the searcher has an easy time believing you built the thing they need. Chapter 5 covers how to find these groups in Keyword Planner. Chapter 8 writes the ads.
All 12 Software Sites We Checked Already Have Pages for One Industry or Use Case. Point the Matching Ad Group There.
The good news for most software companies is that the pages often already exist. We read the sitemaps of the twelve sample sites, and every one lists pages built around a single industry, feature or use case. A few examples:
- Patriot Software has payroll pages for churches, dental offices and healthcare practices.
- Copper has CRM pages for construction companies, agencies and consultants.
- OnTheClock has time-tracking pages for call centers, churches and cleaning businesses.
- When I Work has separate pages for employee scheduling, team messaging and its time clock.
Each of those pages is a ready-made answer to one search. The ad group for "construction crm" should land on the construction page, not the home page, and not a generic features page. Where an ad group has no matching page yet, building one is usually cheaper than paying for the mismatch on every click for the next year.
Two of the twelve sites go further, with landing page templates that strip the full site menu down to a logo and, at most, a button or two. Pages like that keep the visitor on one task, the trial or the demo, instead of offering them the blog, the careers page and the login. Whether they help depends on your product; a buyer comparing tools may want the menu. Test it rather than assume it.
One caution about the ads themselves. An ad's display URL is typed text, not proof of where the ad sends people, so the ads in Google's library can't tell you which page an advertiser actually uses. Check your own final URLs in the account. If most of them end in a slash and nothing else, every ad you run is sending buyers to the same front door.
On a Phone, 10 of the 12 Software Home Pages We Tested Needed Over 4 Seconds to Show Their Main Content. Test Yours Before You Pay for Clicks.
Landing page experience is mostly about whether the page answers the search. But a page nobody waits for can't be useful. We ran each sample site's home page through Google's PageSpeed Insights on a phone setting, one test each. On ten of the twelve, the main content took more than four seconds to appear. Two loaded in under two seconds, so fast is possible.
One test is a snapshot, not a verdict, and a home page shouldn't be your landing page anyway. Run PageSpeed Insights on the pages your ads actually use before you spend on them. If the page is slow, fix the biggest item the report names first, usually images or scripts that load before the content.
Top-of-Page Clicks Go to the Best Ad Rank, and Quality Is the Cheaper Way to Raise It.
Ads in the top slots draw more of the clicks than ads lower down, and software companies bid hard for them. There are two ways to climb: raise the bid, or raise quality. The bid is fast and costs more on every click. Quality is slower and usually makes future clicks cheaper.
In practice you do both, but in order. Build tight ad groups pointed at matching pages first, so Google has a reason to rank you. Then set bids, which Chapter 7 covers, with the impression share columns that show how often you're winning the top of the page.
Next: Most Software Searches We Pulled Name a Brand. Build Your Keyword List From the Ones That Don't.
Grouping only works once the keyword list is right, and most software keyword lists start out full of brand names that belong somewhere else. Chapter 5 builds the keyword list, starting from your competitors' sites and leaving out the brand names, login searches and free-tool searches that soak up software budgets.
Every chapter, from tracking to brand campaigns, competitor campaigns and the rules for free trials, is on the guide's home page. Chapter 3 shows why the click-to-trial rate matters so much; this chapter is how you raise it.
On the software accounts we run through Google Ads management, the page list gets built alongside the ad groups: every group has a page before it goes live. Start with yours: list your ad groups in one column, the pages they point to in the next, and see how many rows end in the home page.
