Analytics on a store is only worth setting up if it changes a decision. Plenty of dashboards are opened once, admired, and never used again because nothing in them suggests an action. The goal is a short set of numbers that answer three questions: where are customers coming from, where do they stop, and what should we change this week?
Decide what you are trying to answer
Write down the questions before you install anything. For most small stores they are: which products deserve more effort, which pages need better information, whether paid spend is producing orders, and whether repeat customers are coming back. Each question suggests a specific measurement. Starting from a list of available charts instead produces a reporting system nobody can use, because there is no decision attached to any of it.
Install tracking properly
Most store platforms can install their own analytics and report sales automatically, which is the sensible starting point. Beyond that, check that purchase, add-to-cart and checkout events fire once and at the right moment, and that they carry sensible values. Test by placing a real order and confirming the numbers, then refund it and see whether the reporting handles that correctly. Broken or duplicated tracking is common and quietly distorts everything downstream.
Respect consent and privacy
If you serve Indian customers, some of whom may be in scope of data protection expectations, be clear about what you collect and why. Load optional analytics and advertising scripts only after consent where that is required, give a genuine choice, and remember the choice. Keep your own cookie and tracking practices documented, and make sure your email and messaging tools follow the same rule. This is a compliance matter and a trust matter, and it is easier to do correctly from the start.
Report on behaviour, not just totals
Total revenue hides more than it shows. Break it down by product, category, source, device and new versus returning customer. Look for pages with views but no add-to-cart, and for products with add-to-cart but no purchase, because those two groups point at different problems. Check how long people spend on a product page, and use the store search report to find naming gaps. Be suspicious of very high traffic with flat sales, since it usually indicates a mismatch between what people expected and what they found.
Build a reporting rhythm that survives
Weekly: sessions, orders, conversion by source, and the checkout drop-off point. Monthly: product performance, search terms, repeat customer share, and whether marketing spend produced orders at a cost you can live with. Keep a written note of what changed, what you expect to happen and by when, so the next month has something to compare against. If a provider sends you a report, ask to see the underlying data yourself rather than relying on a screenshot.
Understand where each number comes from
Store analytics, your payment provider's settlement report and your bank all count slightly different things. Settlements arrive net of fees and sometimes net of refunds in a later cycle, so revenue recorded on the day of purchase will not match the money that arrives. Decide which figure answers which question: use store analytics for product and funnel decisions, and the settlement report for cash. When you report internally, state which source you used, otherwise a disagreement about numbers turns into a disagreement about strategy.
Set up the marketing side too
If you run ads, content or email to your store, connect those tools so you can see which campaign produced orders rather than only clicks. Track the campaign, the landing page and the purchase together, and check that you can import cost data so profit is visible, not only revenue. In many cases this single connection is what makes paid activity accountable. Be honest about attribution limitations: platforms claim credit differently, and last-click reporting is simpler but not the whole story.
Watch the return rate as an outcome
Returns affect your real revenue more than any traffic figure, so treat the return rate by product as a first-class measure. A page that sells well and returns badly is not a success. Pair it with stockout frequency and with the proportion of orders coming from repeat customers, because a store that only ever acquires new customers at full acquisition cost is carrying a risk. None of these measures are interesting on their own; together they tell you whether the business is actually getting healthier.
Keep the measurement affordable
Start with what the platform and free search reporting give you. Add a tool only when you know the question it answers, and check whether the same answer could come from an export you already have. Agree in advance with whoever works on the store whether the fee covers the setup only or also monthly checking and reporting, and keep the underlying data accessible to you so you are never dependent on someone else's summary.
More of the measurement side is in our e-commerce guides. If you want tracking configured correctly as part of a build, our e-commerce development work starts at INR 35,000 per project, with hosting, platform subscription, payment gateway charges, shipping and ad spend billed separately by their own providers, and you can tell us what you currently look at for a straight answer about what would help.
Common questions
Which ecommerce reports should I actually check each week?
Start with five: sessions and where they came from, product and category pages with views but no add-to-cart, the checkout step where most sessions stop, completed orders by product, and the store search report showing what customers typed when they could not find what they wanted. Everything else can wait until you can answer what you will do differently based on those five.
Do I need paid analytics or reporting tools?
Most stores can get a sensible picture from tools included with their platform plus free search reporting. Paid tools help when you have enough traffic to need deeper segmentation, when you sell across several channels, or when you want automated reports instead of exporting data yourself. Buy one only after you know the specific question you need it to answer, because a dashboard nobody reads changes nothing.
Why do my numbers not match my sales figures?
Usually because two systems count different things. Bank settlements, the platform order report and analytics will each treat refunds, cancellations, taxes and shipping differently, and a cancelled order may still appear as a conversion. Decide which single source is authoritative for revenue and state it in your reporting, then reconcile the others to it monthly instead of expecting them to agree automatically.