Every month, money goes out: Google ads, maybe Facebook, the website, truck wraps, a mailer, a sponsorship for the little league team. Every month, jobs come in. The question that should be easy and never is: which of those dollars brought which of those jobs?
The marketing industry's answer is "attribution," and the honest truth about attribution is that perfect attribution does not exist, not for you, not for companies with eight-figure budgets and a floor full of analysts. The customer journey is too messy to fully reconstruct. But, and this is the actual point of this post, you don't need perfect. You need directional truth: enough signal to confidently spend more on what's working and kill what isn't. That's achievable for a normal business with a few simple disciplines and no new software degrees.
Why this is genuinely hard
Picture a real customer. She sees your wrapped truck at a stoplight in March. In April, a neighbor mentions you in a Facebook group. In May, her AC dies, she searches your name, clicks your Google Business Profile, reads reviews, clicks through to your website, leaves, comes back two days later from a Google ad because you were retargeting her, and calls.
Which marketing gets credit? The truck wrap that planted the name? The neighbor you don't control? The ad that happened to catch the last click? Software will confidently answer "the ad," and that answer is wrong in every way that matters.
That's the whole problem in one story. Most measurement systems credit the last touch before the sale, while most of the persuasion happened earlier, in places that are hard or impossible to measure.
Last-click: useful, and systematically biased
Last-click attribution gives 100 percent of the credit to the final click before the conversion. It's the default lens of most reporting, and you should understand its bias rather than discard it.
Last-click systematically over-credits the bottom of the funnel: brand searches, your Google Business Profile, retargeting ads. These harvest demand that already existed. It systematically under-credits everything that created the demand: the wrap, the yard signs, the reviews, the word of mouth, the billboard. Cut the demand creators because "the report says they don't convert" and a few months later the harvesters mysteriously dry up too.
Modern analytics tools try to spread credit across touches with data-driven models; Google's documentation on attribution in Analytics explains the options if you want the deeper dive. Worth knowing, but for a business doing dozens of jobs a month rather than thousands, fancier models mostly reshuffle small numbers. Your edge comes from somewhere simpler.
The four disciplines that get you 80 percent
1. Ask "how did you hear about us," and actually write it down
The least glamorous tool is the most powerful one, and almost nobody does it well. Three failure modes to avoid:
- It doesn't get asked. Whoever answers the phone is busy. Make it part of the script, every call, every form follow-up, no exceptions.
- It gets asked badly. "Google" is not an answer; it's a category. The follow-up question is the discipline: "Were you searching for the kind of work, or did you already know our name?" The first means search and SEO are working. The second means something else, a referral, a truck, a sign, created the demand and Google just looked up the number. Those are completely different marketing wins.
- It doesn't get written down. The answer has to land somewhere permanent next to the job record, not in the wind. A column in a spreadsheet is plenty.
Self-reported answers are imperfect; people misremember. But asked consistently for six months, this one habit beats most software, because it captures the touches software can't see: referrals, reputation, the neighbor in the Facebook group.
2. Call tracking, used carefully
Call tracking numbers let you put a unique phone number on each channel, the ads, the mailer, the website, and know exactly which channel made the phone ring. For phone-heavy service businesses, this closes the biggest measurement hole, since calls are usually most of the leads and the part online analytics sees worst.
Two cautions. First, never let tracking numbers contaminate your core business listings: your Google Business Profile and major directories should carry your one true number consistently, because conflicting numbers across listings can undermine the local presence you've built. Use tracking numbers on ads and campaigns, not on your permanent identity. Second, if calls are recorded for quality, say so and follow the rules; recording laws vary by state.
A side benefit owners underrate: listening to a sample of tracked calls tells you how many "leads" were real, and how well the phone is being answered. Plenty of marketing problems turn out to be phone-answering problems.
3. Tag your links, separate your destinations
When you control the link, label it. UTM tags, the little labels added to the end of a URL, tell your analytics exactly which campaign a visitor came from, and they cost nothing but consistency. For offline campaigns, use a distinct destination: the mailer points to a specific page, the QR code on the truck to another. If a campaign's traffic can't be told apart from everything else, you've already decided not to measure it.
4. Read the free scoreboards you already have
Your Google Business Profile reports calls, direction requests, and website clicks it generated; the Business Profile help center covers where to find them. Your ad platforms report their own numbers, with a grain of salt since they're grading their own homework. None of these is the whole truth. Together with your "how did you hear about us" column, they triangulate.
Directional truth over perfect data
Here's the mindset shift that makes all of this manageable. You are not trying to compute the exact ROI of every dollar. You're trying to sort your channels into three buckets:
- Obviously working. Leads consistently name it, cost per booked job is comfortably below what a job is worth. Feed it more budget.
- Obviously failing. Months of spend, almost nothing attributable by any method. Kill it without sentiment.
- Unclear. Either give it a fair, time-boxed test, or consciously fund it as a brand bet, the truck wrap may never win a report and still be worth every penny.
Two rules keep the buckets honest. Judge on booked jobs and revenue, not clicks or impressions; a channel that produces cheap leads who never buy is a failing channel with good stats. And judge monthly or quarterly, never weekly; at small-business volumes, one week is noise, and reacting to noise is how good channels get killed and bad ones get excused.
A scorecard you'll actually keep
One page, once a month, per channel:
- Spend: what it cost this month, including your time if it's significant.
- Leads: how many, by tracking number, tagged link, or the phone question.
- Booked jobs: how many became revenue. The number that matters.
- Revenue: dollars those jobs brought in.
- Cost per booked job: spend divided by booked jobs. The decision number.
Ten lines in a spreadsheet. After three months you'll see shape; after six you'll make budget decisions with a confidence most of your competitors never reach. The SBA's guidance on marketing and sales makes the same underlying point: marketing is a budget line that should answer for itself like any other.
One practical tip for keeping the habit alive: put the scorecard review on the calendar, same day every month, fifteen minutes, right after you reconcile the books. Attribution systems don't die from bad math; they die from skipped months. The owner who reviews mediocre data every month beats the owner with a perfect dashboard nobody opens.
The traps on the way out
- Double counting. The ad platform claims the lead, the SEO report claims it, and the customer says "my neighbor." Pick one source of truth for booked jobs, we'd argue for the phone question plus your job records, and treat platform reports as supporting evidence.
- Vanity metrics. Impressions and reach measure exposure, not money. Useful context, never the verdict.
- Cutting the demand creators. The last-click bias again. Before killing a channel that "never converts," ask whether your brand-name searches drop when it stops.
- Reports that grade their own homework. Any vendor reporting on their own channel deserves your scorecard as a cross-check, not blind trust.
This is also exactly the kind of operating discipline we work through with owners in our Command Advisor engagements, because the businesses that grow aren't the ones with perfect data, they're the ones that look at directionally honest data every single month.
Want a website that makes this easy?
Measurement starts with a site built to capture and label every lead. Omnyra builds done-with-you websites live on a call: first draft in 24 hours, live in 7 days, guaranteed, with 1,500+ small business sites built in the last 90 days. Tiers start at $500, pay-in-4 and Klarna available, and the Max tier connects your site and phones to Jobber, ServiceTitan, or GoHighLevel, so every lead lands tagged with where it came from. Veteran-owned, Wilmington, NC.