Smarter Analytics for Small Teams: Fewer Numbers, Better Decisions

Ask a small business owner how the website is doing and you will usually get a sessions number. Ask what they changed because of it and the conversation stops. That gap, between a metric that is reported and a metric that causes a decision, is the entire problem with small business analytics. It is not a tooling problem, and buying a better dashboard will not fix it.

Small teams have a real advantage here that enterprises do not: nobody has to be persuaded. There is no analytics committee, no quarterly instrumentation roadmap. If the owner and one marketer agree that a number matters, it can be tracked by Friday. The constraint is attention, not budget, so the design goal is a small number of metrics that people actually look at.

The test for whether a metric is worth tracking

Before adding anything to a dashboard, answer one question: what would you do differently if this number moved 20% in either direction? If you cannot name the action, the metric is decoration. Applied honestly, this deletes most of what sits on a typical small business dashboard.

A useful shorthand: if a number can go up while your business gets worse, it is a vanity metric. Traffic can spike from a post that brings the wrong audience. Follower counts can climb from a giveaway that attracts people who will never buy. Qualified leads and cost per qualified lead do not have that property, which is why they earn dashboard space.

Four numbers are usually enough

For most small businesses we work with, the reporting set that survives contact with reality is roughly four metrics, reviewed weekly, with everything else available on request but not on the wall.

1. Qualified demand: leads, bookings, or orders that meet a definition you wrote down. The written definition matters more than the number.
2. Cost to acquire that demand, split by channel, including the labor hours nobody counts.
3. Conversion rate at the single weakest step of your funnel, whichever step that currently is.
4. Retention or repeat rate on a fixed window, because acquisition economics are meaningless without it.

The GA4 setting that deletes your history

This one is worth a section on its own because it catches almost everybody. Google Analytics 4 lets you set how long user-level and event-level data is retained, and the standard options are 2 months or 14 months. Properties are frequently left on the shorter setting, and nobody notices until the day they try to compare this quarter against the same quarter last year.

The subtlety is that the setting does not affect standard aggregated reports, which keep working across long date ranges. It affects Explorations and funnel analysis, which are exactly the tools you reach for when you finally have a real question. So the failure is delayed and silent: everything looks fine for a year, then the one analysis you actually needed comes back empty for the earlier period.

• Open Admin, then Data settings, then Data retention, and set event data retention to 14 months, the maximum on the free tier.
• Enable the option to reset user data retention on new activity, so returning users do not expire mid-analysis.
• Do this today rather than when you need it. The setting is not retroactive; data already aged out does not come back.
• If you need multi-year trend analysis, export to BigQuery, which has no such window.

Annotate, or you will misread your own charts

Six months from now, a traffic cliff in March will be a mystery. Was it a Google update, a broken tag, a paused campaign, or the week the site went down? Nobody remembers, and the instinct will be to invent a plausible story. Keep a running log of launches, campaign changes, price changes, outages, and site migrations with dates. It takes a minute per entry and it is the difference between reading a chart and guessing at one.

This applies to search performance especially. If you are making changes to site structure or internal links, log them. Attribution for an SEO improvement is nearly impossible after the fact without a timeline, and internal linking changes in particular tend to show up weeks after they ship.

Attribution: be honest about what you cannot know

Small businesses are sold attribution models with more confidence than the underlying data supports. Between cookie restrictions, consent banners, ad blockers, cross-device journeys, and people who hear about you from a friend and type your name into Google, a meaningful share of your conversions are unattributable. Pretending otherwise leads to defunding the channels that create demand in favor of the channels that harvest it.

The cheapest correction is a single optional question on your enquiry form: "How did you hear about us?" Free text, no dropdown. The answers are messy and self-reported, and they consistently surface sources your analytics never sees. Read them against your platform data rather than instead of it. Where they disagree badly, trust neither and run a holdout test.

Build the smallest thing that gets read

A one-page weekly summary that the team actually reads beats a real-time dashboard nobody opens. We usually land on a short recurring email or a single pinned page: the four numbers, week over week, plus two sentences on what changed and one decision made. The two sentences are the valuable part. A dashboard with no narrative gets skimmed; a narrative with numbers attached gets discussed.

Written by

Valon Badivuku

Digital Strategist

Valon Badivuku is a Digital Strategist at ThisCom, helping brands get seen and become visible online through strategies that turn attention into lasting growth.

All articles by Valon Badivuku →