You need about eight numbers to know whether your marketing is working. Not a dashboard. Not a weekly report you dread. Eight numbers, checked monthly.
The marketing industry produces metrics the way a lawn produces grass, and most of them exist to make reports look thorough. This article is the short list: what’s worth tracking for a small business, what each number actually tells you, and how to read them together without making it a hobby.
Why this matters
The problem isn’t too little data. It’s that the data you’re handed measures the wrong thing.
- Vanity metrics reward the vendor, not you. Impressions, reach, followers, and likes all go up and to the right while the phone stays quiet. They measure activity. You’re paying for customers.
- What you can’t measure, you can’t cancel. Roughly 70% of small business owners can’t say what their marketing provider actually does each month. Without your own numbers, every renewal decision is a trust exercise.
- Gut feel has a lag problem. By the time a slow month is obvious in the bank account, the cause is three months old. The right numbers surface trouble while it’s still cheap to fix.
- Money decisions need money numbers. “Engagement is up 40%” can’t tell you whether to renew a $500 monthly service. “That channel produced nine calls and two jobs” can.
The test for any metric: could it change a decision about where your money goes? If not, it’s decoration.
The metrics that matter
Marketers sort these into what happens before the contact and what happens after. Plainly: how people find you, and what it turns into.
Leads. How many people contacted you this month: calls, form fills, texts, walk-ins. The single most important marketing number you have. If you track nothing else, track this by hand in a notebook.
Lead source. For each lead, where it came from. Marketers call this attribution, which means figuring out which effort produced the call. The reliable version for a small business is asking “how did you find us?” and writing the answer down.
Website visitors. How many people reached your site, and roughly from where: Google search, ads, social, or typing your name directly. Direction matters more than precision here; a steady climb means growing visibility.
Conversion rate. Of the people who visited, how many contacted you. This is the number that tells you whether your website works. A hundred visitors and one call is a website problem, not a traffic problem, and more visitors won’t fix it.
Cost per lead. What you spent on a channel divided by the leads it produced. This is the number that makes channels comparable: the networking breakfast, the ad campaign, and the SEO retainer can finally be judged side by side.
Close rate. How many leads became paying customers. If lots of people call and few buy, the marketing may be fine and the problem lives in pricing, follow-up, or who the marketing is attracting.
Customer value. What a typical customer is worth to you, including repeat work. You need this once, not monthly, because it sets the ceiling on what a lead is worth paying for. A $6,000 customer makes a $60 lead a bargain; a $80 customer doesn’t.
Reviews. Count and average rating, yours and your closest competitors’. It’s the scoreboard most local buyers check before calling anyone.
Notice what’s missing: followers, impressions, reach, likes, email opens. Fine as diagnostics when you’re studying a specific channel. None of them belongs in the monthly eight.
Putting it to work
- Set up the world’s smallest tracking system. A spreadsheet with one row per month and one column per metric. The prompt below builds it and defines each column. Free analytics tools can feed it, but the spreadsheet is the system.
- Ask every lead how they found you. Train yourself and anyone who answers the phone. It’s imperfect, and it’s still the best attribution a small business can buy at the price of one question.
- Read trends, not months. One bad month is weather. Three months moving the same direction is a trend worth acting on. Compare against the same season last year before panicking; most local businesses breathe with the calendar.
- Judge each channel against its own clock. Ads should show leads within weeks. Search and content take months to move, which is normal and why your strategy gave each channel a minimum term. The metrics tell you what happened; the strategy tells you when judging is fair.
- Let the numbers ask questions, not answer them. A falling conversion rate doesn’t say what broke; it says look at the website. Metrics point; you investigate.
One honest limit. Small numbers are noisy. When you get fifteen leads a month, one chatty referral customer can swing a percentage badly. Read direction over quarters, hold the numbers loosely, and never let a single month’s dip stampede you into scrapping a plan that’s on schedule.
If you want the deeper treatment of analytics tools and how they work, the analytics and data-driven marketing guide covers it. This article is the part you’ll actually use monthly.
Take this prompt with you
Copy this into ChatGPT, Claude, or Gemini and answer its questions. It builds your tracker and then reads it with you each month.
Help me set up simple marketing metrics tracking for my small business, then help me read it each month.
My business: [what you do, where, roughly how many customers you get per month, and what a typical customer is worth including repeat work]
My marketing channels right now: [list what you do: e.g. website, Google Business Profile, ads, social, networking, referrals]
Step 1. Build my tracker. Give me a table I can copy into a spreadsheet with one row per month and columns for: leads, leads by source (one column per channel I listed), website visitors, conversion rate, cost per lead by paid channel, close rate, and review count and average rating. For each column, tell me in one line where the number comes from and how to get it in under five minutes.
Step 2. Set my baselines. From what I've told you, estimate what healthy looks like for a business like mine for each metric, and mark which estimates are rough so I know to refine them with my own data.
Step 3. Monthly reading. Each month I'll paste the updated table and you'll tell me: what moved meaningfully versus normal noise given numbers my size, the single most likely cause worth checking for anything that moved, and the one action you'd take next month. Flag any metric that's moved the same direction three months running.
Rules: keep everything in plain language, compare month-to-month trends rather than judging any single month, account for seasonality by asking me which months are naturally busy and slow, and never recommend tracking more numbers than this list unless I ask.
