Category

Analytics and dashboards

Most small businesses have more marketing data than they can use and less information than they need. This category covers the short list of numbers worth seeing every week, how to set a baseline before you change anything, and how to read a monthly report from an agency critically. It spends real time on the problem that ruins small business analytics, which is that at low volumes almost every week-to-week move is noise. After reading it you should have a review habit you can actually keep, and the ability to tell when a number has genuinely changed.

Start here

• The numbers a small business owner should see every week Read this first. A short list you can actually keep up with beats a dashboard nobody opens.
• Setting a baseline before you change anything Without a baseline you cannot prove anything worked, and this is the cheapest step to skip and regret.
• Small sample sizes and how to avoid fooling yourself The trap that makes local businesses chase noise for months. Read it before you act on a weekly change.

All twenty articles on analytics and dashboards

In reading order rather than by date. 20 published so far, and the rest are written and scheduled. Titles without a link are not live yet.

1. The numbers a small business owner should see every week

   Five numbers, and no more: revenue against the same week last year, new leads or bookings, spend on your largest channel, one operating cost you can change this week, and cash on hand.

2. Google Analytics 4 for people who do not want to learn Analytics

   You need three things from GA4 and you can ignore the rest of it forever: how many people arrived, where they came from, and how many did the thing you care about.

3. What a business dashboard should actually show

   A dashboard should show the numbers that change what you do, labelled with how confident you can be in each one.

4. Vanity metrics and the ones that predict revenue

   A vanity metric goes up when someone does work. A predictive metric goes up shortly before money arrives.

5. Setting a baseline before you change anything

   A baseline is a dated record of where every number stood before you changed something, captured in a form you cannot revise later.

6. Connecting your POS data to something useful

   Export daily sales, transaction count, and daypart from your point of sale, put it next to your marketing data on a shared date column, and stop there.

7. Tracking phone calls as conversions

   Use a call tracking number that swaps by traffic source, count only calls over a duration threshold, and record answered versus missed separately.

8. UTM tags explained for small business owners

   UTM tags are extra text you add to the end of a link so your analytics can tell where a visitor came from.

9. Attribution when you have five marketing channels and no budget

   Ask every customer how they found you, log the answer, and compare that to your tagged digital data.

10. Why your website analytics and your ad platform disagree

    They disagree because they are counting different things, over different windows, with different rules about who gets credit.

11. What to measure in the first 90 days of any marketing engagement

    Measure setup completion and leading indicators, not revenue. In ninety days you can honestly assess whether tracking is live, whether the work shipped, whether impressions and.

12. Reading a monthly marketing report critically

    Check four things in this order: what period is being compared, what the base numbers are behind every percentage, which metrics are counted versus estimated, and what is missing that.

13. Google Search Console: the free tool most businesses ignore

    Search Console is Google telling you, for free, which searches showed your site, how often, where you ranked, and who clicked.

14. Measuring AI search visibility honestly

    You can count AI crawler visits in your server logs, and you can estimate mention rates by asking the same question many times and recording the spread.

15. Server log analysis for AI crawler traffic

    Your server records every request to your site, including every fetch by an AI crawler.

16. Small sample sizes and how to avoid fooling yourself

    A metric moving on a small base is usually noise. Eleven leads after eight is three leads, not a thirty-eight percent improvement.

17. Seasonality and how it distorts month-over-month comparisons

    Compare each month to the same month last year, not to the month before it.

18. When a metric moves, telling signal from noise

    Run four checks before you react. How large is the base, how far outside the normal range is the move, did anything change in the tracking, and did anything change in the world.

19. Building a monthly review habit that survives busy season

    Make it thirty minutes, on a fixed date, with a one page agenda and a named owner.

20. What data you should own and be able to take with you

    Own the domain, the website files and database, the analytics and Search Console properties, the ad accounts, the Google Business Profile, your customer list, and your historical.

Be honest with yourself

When this category is not your problem

If you already know which channel brings your customers and you act on it, more measurement is procrastination. Dashboards feel like progress and are not the same thing.

And if your traffic is genuinely small, most of what you can measure will not reach significance in any useful window. Track a few things monthly, act on judgment, and stop refreshing.

Where to go next

• Website conversion Measurement is what tells you whether a conversion change did anything at all.
• Google Ads and paid local Paid spend is where bad measurement costs money fastest.
• AI search and AEO AI visibility is mostly a measurement problem, and a badly behaved one.

If you get a monthly marketing report and are not sure what it is telling you, send one to eric@seod.com with the client details removed and I will tell you which numbers are real, which are selected, and what is missing. I do this for people who are not clients.

A chart is not a decision, which is the whole premise of Dashboards & data: what changed this month, why it changed, and what to do next.