Start smaller than you think
Sellers new to analytics often try to track everything a dashboard tool offers and end up tracking nothing consistently. A short list of metrics checked reliably every week beats a comprehensive dashboard checked once a month.
The underlying reason this works is that a metric only creates value if it changes a decision. A dashboard with forty metrics, most of which no one looks at closely enough to act on, provides less real decision-making value than five metrics that someone actually reviews every week and adjusts pricing, ad spend, or inventory in response to.
The starter list
Revenue and unit sales, tracked weekly, to spot trend direction early. Conversion rate (sessions or clicks-to-purchase), because it's the single number most sensitive to listing quality, price competitiveness, and review count — see Conversion Rate: What's Good, and How to Improve It. Contribution margin per unit or per order (not just revenue — see the Unit Economics Calculator), because revenue growth without margin discipline is a common trap for growing sellers. Return rate, because a rising return rate is often the earliest signal of a product quality or listing-accuracy problem, before it shows up anywhere else. Advertising efficiency (ACOS/ROAS/TACOS — see ROAS vs. ACOS vs. TACOS) if you're running ads.
Why this specific set, and not a different one
Each metric on this list answers a distinct question that the others can't answer for it. Revenue tells you if the business is growing. Conversion rate tells you if the listing itself is working, independent of how much traffic you're getting — a listing can have flat revenue because traffic dropped even though conversion rate improved, and conflating the two hides that. Contribution margin tells you if growth is actually profitable, since revenue and profit can move in opposite directions if costs (advertising, returns, fees) are creeping up faster than sales. Return rate is a leading indicator — it often moves before revenue does, giving you a earlier warning than waiting for sales to visibly decline. Advertising efficiency tells you whether paid growth is sustainable or subsidized.
A worked example: reading the starter dashboard together, not in isolation
Imagine a seller's week-over-week numbers show: revenue up 8%, conversion rate flat, contribution margin per unit down 6%, return rate up slightly, ACOS up meaningfully. Read individually, "revenue up 8%" looks like good news. Read together, the story is different: the growth was likely bought with more aggressive advertising (ACOS up) rather than earned through better conversion, and it's compressing margin — meanwhile the slight return-rate uptick is worth watching in case it's an early signal of a product or listing issue rather than noise. This is exactly why a starter dashboard needs at least these five numbers together, not just a revenue chart.
Marketplace-specific additions
Once you're established on a specific marketplace, add that platform's native performance metrics — see Marketplace-Specific KPIs, Explained for what Amazon's Unit Session Percentage, Buy Box win rate, and similar metrics actually measure.
Where to keep this dashboard
A simple spreadsheet updated weekly is entirely sufficient at low order volume — see the KPI Dashboard Template for a starting structure. Dedicated analytics tooling becomes worthwhile once manual tracking becomes a meaningful time cost, not before.
Adding metrics over time
Resist the urge to add a new metric to your dashboard just because it's available. A reasonable trigger for adding one: you've noticed a recurring question you can't answer with your current metrics ("is my return rate different by channel?" → add channel-split return rate), not simply "this tool also shows me X." Every metric added should replace idle curiosity with something you'll actually check weekly and act on.
The mistake to avoid
Chasing a vanity metric (like raw traffic or impressions) that doesn't connect to revenue or margin. Every metric on this starter list ties directly to a business outcome — sales, profitability, or customer satisfaction — rather than an intermediate number that only matters if it moves one of those.
Best practices
- Review all five starter metrics together in the same sitting, not one at a time in isolation.
- Add a new metric only once you have a specific recurring question your current metrics can't answer.
- Log the week's numbers somewhere durable (a spreadsheet row), not just glanced at and forgotten.
- Treat a rising return rate as an early-warning signal worth investigating immediately, not something to wait out.
Troubleshooting
Revenue is up but I have a nagging feeling the business isn't actually healthier. Check contribution margin and advertising efficiency for the same period — revenue growth funded by margin compression or rising ad spend is a common pattern that a revenue-only view hides completely.
My conversion rate looks fine but sales are flat. Check traffic — a flat conversion rate with flat sales usually means traffic itself is flat or declining, which is a different problem (visibility, ranking, ad reach) than a listing-quality problem.
I don't have enough order volume yet for these numbers to feel meaningful. At very low volume, a handful of orders can swing a percentage-based metric wildly week to week — track raw counts (units, orders) alongside the percentages until volume is high enough for the rates to stabilize.
FAQs
How many metrics is too many for a starter dashboard? There's no fixed number, but if you can't recall roughly what each metric on your dashboard measures and why it's there without looking it up, that's a sign you've added more than you're actually using.
Should I track metrics daily or weekly at this stage? Weekly is usually the right cadence for a new seller — daily data is noisier and encourages overreacting to normal variance, while weekly smooths enough of that noise to show real trend.
What's the very first metric to start tracking if I can only pick one? Contribution margin per unit — it's the metric most likely to reveal a business model problem (pricing too low, costs too high) that revenue or conversion rate alone would mask.