Measuring Calls, Leads and Sales

Updated: November 13, 2026 · Leer este capítulo en español

To measure Google Ads honestly, you have to follow the money past the click: to the call, the form, the appointment and, finally, the paying customer. Google will happily show you impressions, clicks and click-through rates, but none of those pays rent. This chapter explains what should count as a conversion, how to track and listen to calls, the difference between cost per lead and cost per customer, how to connect sales back to your ads, which numbers to review every week and month, and how to decide when a campaign should stop, be fixed or grow.

How to measure Google Ads: what counts as a conversion

In Google Ads, a conversion is whatever you tell Google is valuable. Google's own page About conversion measurement puts it simply: you choose what you identify as valuable, such as a purchase, a sign-up or a phone call. That freedom is the trap. Google will count whatever you choose, report it proudly, and, if you use smart bidding, spend your budget to get more of it.

The setup itself belongs in Before You Spend a Dollar. This chapter is about judgment: deciding what deserves to be counted. A simple test helps. For each action, ask: "Would I pay money for one more of these?"

Google Ads lets you mark conversion actions as primary (used for bidding and in the main Conversions column) or secondary (recorded for observation only). Keep the soft signals as secondary. If you are using smart bidding, that choice is not cosmetic: Budgets and Bidding explains why the system optimizes for exactly what you count.

Tracking calls, and listening to them

For an HVAC company or a dental office, most leads arrive by phone, so call tracking is not optional. Google Ads can track calls in a few ways, described in its page About phone call conversion tracking: calls made directly from your ads (through call assets), calls to a Google forwarding number shown on your website to people who clicked an ad, and taps on the phone number on your mobile site. For the first two, you set a minimum call length, and only calls that last at least that long count as conversions.

Choose that minimum with care. Too short, and wrong numbers and robocalls count as leads. Too long, and a quick "Can you come today at 3?" that turned into a job is ignored. Look at your real calls: how long does a genuine first conversation usually take at your business?

Many businesses also use a separate call tracking service that assigns numbers by source and records calls. Recordings are the best quality-control tool in local advertising, because they show what the numbers cannot: whether the caller wanted what you sell, whether they spoke Spanish and got an English-only answer, whether the person on the phone gave a price and lost them.

Recording has a legal side, and Florida is strict. Section 934.03 of the Florida Statutes generally makes recording a call lawful when all parties have given prior consent, which is why recorded business lines here usually begin with a notice such as "this call may be recorded." Set up that notice before you record anything, and ask your attorney about your specific situation; this book is not legal advice. Note also that some platforms limit recordings in sensitive categories: Google's Local Services Ads, for example, do not offer call recordings for health care businesses in the US (see Local Services Ads and Google Maps).

What pros do: they listen to a sample of calls every month, from the start of the call to the end, with a simple checklist: Was it a real customer? Was the service something we do? Did we answer in time and in their language? Did we book it? Ten calls a month will teach you more about your campaign than any dashboard.

Cost per lead vs cost per customer

Most reports stop at cost per lead (CPL): what you spent divided by the number of conversions. It is useful, but it is a halfway number. Leads are not equal. Ten calls from people asking whether you sell window units, and two calls from homeowners whose central AC died, are twelve "leads" with very different value.

The number that tells you whether ads pay is cost per customer, sometimes called cost per acquisition: what you spent divided by the number of leads that became paying customers. Here is an illustrative month for our example HVAC company in Hialeah; the figures are invented to show the method:

A $40 lead sounds cheap, and a $167 customer sounds expensive, but neither means anything until you compare it with what a customer is worth after the cost of doing the job. That comparison is the heart of the budget method in Budgets and Bidding, and it is the only honest answer to "are my ads working?"

Following each lead to a sale

To get cost per customer, you need a record of what happened to each lead. A CRM does this well, but a shared spreadsheet is enough to start. What matters is that someone fills it in every day, not once a month from memory.

• Date and time of the call, form or message.
• Source: Google Ads, Local Services Ads, Maps, referral, repeat customer. Ask "How did you find us?" and also use tracking numbers, since people often misremember.
• Name and phone or email.
• Service requested, and whether it is one you offer.
• Language of the conversation.
• Outcome: booked, quoted, lost (with a reason: price, timing, no answer, out of area), or not a real lead.
• Sale amount when it closes.

For our example dental office in Doral, the sheet might reveal that implant consultations from ads close well, but cleaning inquiries mostly ask whether a specific insurance plan is accepted. That is not something Google can tell you. It might lead the office to state accepted plans on the landing page and spend less on general searches.

Sending sales back to Google: offline conversions

Once you know which leads became customers, you can tell Google. This is called offline conversion import, and Google describes it in Set up offline conversions using Google Click ID. The concept is straightforward:

1. When someone clicks your ad, Google adds an identifier to the page address, the GCLID (Google Click ID).
2. Your form or CRM stores that identifier next to the lead.
3. When the lead becomes a customer, you upload the sale with its identifier, by file or through an integration.
4. Google connects the sale to the click, the keyword and the campaign.

Google also offers enhanced conversions for leads, which matches sales using lead information such as an email address, sent in hashed (scrambled) form, in addition to the click identifier. According to Google's conversion help pages, click identifiers must be uploaded within 90 days of the click.

Why bother? Because it changes what smart bidding chases. A campaign optimized for form fills learns to get form fills. A campaign that also learns which forms turned into paying jobs can lean toward the searches that produce real customers. This is more advanced work and needs some technical setup, so it usually comes after the basics are reliable. The technical plumbing also changes over time; Google's help pages already note changes to the technical route these uploads take, which is one more reason to describe the goal here and leave the click-by-click setup to the current documentation.

The numbers to review every week and month

You do not need fifty metrics. You need a short list, reviewed on a fixed day, written down so you can compare.

Clicks, impressions and click-through rate still matter, but as diagnostics. If calls drop, they help you find out why. They are not the result.

Attribution: why the numbers never match

Sooner or later, your spreadsheet and Google Ads will disagree. That is normal, and understanding why saves a lot of frustration.

• Different dates. Google attributes conversions to the date of the ad click, not the day the call or sale happened. A click on the 30th and a booking on the 2nd land in different months in each report.
• Different credit rules. A customer may click two of your ads before calling. Google's page About attribution models explains that data-driven attribution, which spreads credit across the ad interactions along the way, is the default for most conversion actions, while "last click" gives everything to the final click. At the time of writing, Google no longer supports the older first click, linear, time decay and position-based models.
• Things Google cannot see. The customer who saw your ad, did not click, and called a week later after a friend mentioned you. Or the one who found you on Maps.
• Cross-device paths. People often click on one device and contact you from another. Google reports cross-device conversions in its "All conversions" column, which is why that column can be larger than the main one.

The practical answer is not to chase a perfect match. Pick your own sheet as the source of truth for customers and revenue, use Google's numbers to steer inside the platform, and watch whether both move in the same direction month after month.

When to stop, fix or scale a campaign

Measurement only matters if it leads to decisions. Here is a simple framework, assuming tracking works and the campaign has had enough time and budget to produce real data (a few weeks of clicks is rarely enough to judge).

• Stop or pause a campaign, keyword or service when its cost per customer stays above what a customer is worth, after you have already tried fixing the obvious problems. Stopping is a decision, not a failure.
• Fix first when the leads are real but they are not closing (look at phone handling, follow-up speed, pricing), or when conversions are mostly junk (look at search terms, location settings and the minimum call length).
• Hold when results are acceptable but unstable. More data beats more changes.
• Scale what has stayed profitable for several weeks: raise its budget gradually, add closely related services or nearby ZIP codes, and keep checking that cost per customer holds as you grow.

Our example Cuban restaurant would use the same logic with different numbers: orders, reservations and catering requests instead of service calls. If a campaign brings reservations on weekends that are already full, it is not adding much, however good the conversion count looks.

The final chapter, Doing It Yourself or Hiring Help, covers what a good monthly report should contain if someone else manages your ads. And if you want help setting up this kind of tracking from the start, that is part of our PPC management in Miami. Leads that are not ready to buy today are worth keeping, too; our email marketing services page explains how a simple follow-up list keeps them from going cold.