Product analytics and revenue

Record the actions that represent progress in your product, then measure how often they happen and what they are worth. This workflow follows checkout activity for a shop.

Examples use an engine already open for your application's lifetime and the ZoneTree.Analytics namespace. Every example uses the site shop; use your own site consistently in collection and reporting.

Define a small event vocabulary

Event	Record it when	Context
checkout_started	The customer starts checkout	Customer ID and selected plan
purchase	The order is successfully completed	The same customer ID, plan, and revenue_usd

Use a trusted application ID for VisitorId. It lets the engine recognize the same customer across actions and days without storing the original ID.

Keep units explicit. This example records revenue in US dollars; a report must not add amounts in different currencies as if they were interchangeable.

Record the completed actions

const string customerId = "customer-42";

await analytics.TrackAsync(new AnalyticsEvent("checkout_started")
{
    Site = "shop",
    VisitorId = customerId,
    Dimensions = { ["plan"] = "pro" },
});

await analytics.TrackAsync(new AnalyticsEvent("purchase")
{
    Site = "shop",
    VisitorId = customerId,
    Dimensions = { ["plan"] = "pro" },
    Values = { ["revenue_usd"] = 49.0 },
});

Replace the sample customer and amounts with values from your application. Emit a purchase from the business operation that confirms the order, rather than every time a confirmation page is rendered.

Events are counted as submitted. A visitor ID joins activity; it does not deduplicate purchases. If an integration retries the same business operation, prevent duplicate tracking at that integration boundary.

Compare purchases and revenue by plan

var revenue = Metric.Sum("revenue_usd");
var averageOrder = revenue / Metric.Events;
var sales = await analytics.QueryAsync(new AnalyticsQuery
{
    Site = "shop",
    Event = "purchase",
    Range = DateRange.LastDays(30),
    GroupBy = { "plan" },
    Metrics = { Metric.Events, Metric.Visitors, revenue, averageOrder },
    OrderBy = { Order.By(revenue) },
});

foreach (var row in sales.Rows)
    Console.WriteLine($"{row["plan"]}: {row[Metric.Events]} purchases, " +
        $"{row[Metric.Visitors]} buyers, USD {row[revenue]:F2} revenue, " +
        $"USD {row[averageOrder]:F2} average order");

Console.WriteLine($"Total buyers: {sales.Totals[Metric.Visitors]}");

The report answers four different questions:

Metric	Meaning
Events	Completed purchases
Visitors	Distinct customers who purchased
Sum(revenue_usd)	Revenue recorded on those purchases
Revenue / Events	Average revenue per purchase

One customer buying twice contributes two purchases and one buyer. Overall distinct buyers come from Totals, because a customer can appear under more than one plan.

Measure checkout completion

var checkout = await analytics.QueryAsync(new FunnelQuery
{
    Site = "shop",
    Range = DateRange.LastDays(30),
    Window = TimeSpan.FromDays(1),
    Steps =
    {
        FunnelStep.Goal("checkout_started"),
        FunnelStep.Goal("purchase"),
    },
});

Console.WriteLine($"Checkout completion: {checkout.ConversionRate:P1}");
foreach (var step in checkout.Steps)
    Console.WriteLine($"{step.Name}: {step.Visitors} customers");

This funnel counts customers who start checkout and later purchase within one day. Other actions may occur between the two steps. It measures completion of that path, rather than dividing unrelated event totals.

A one-day window can span sessions. The explicit customer ID keeps the identity consistent across those sessions. Put plan-specific conditions on the relevant step when another step does not carry that dimension.

Prepare recurring reports

For a frequently used plan report, add its rollup before opening the engine:

options.Rollups.Definitions.Add(new RollupDefinition("plan")
{
    Events = ["purchase"],
});

The purchase report can then use pre-aggregated plan rows for eligible ranges. A new ordinary definition starts coverage at activation; it does not rebuild existing purchases automatically.

The checkout funnel still needs retained event history. Keep raw retention at least as long as the periods you intend to investigate. Check each result's plan for coverage warnings.