Analytics & CRO for Ecommerce & D2C Brands

For a small number of recurring reasons, and most can be found in a day. The purchase event may fire twice when a customer reloads the thank-you page. A payment gateway may return the customer to the store on a different address, so the session is cut in two and the order is credited to direct traffic. Orders that were cancelled or refunded remain in analytics. Some purchases are invisible because the visitor declined tracking or used a browser that blocks it.

Perfect agreement is not realistic, and we say so. Analytics samples what browsers allow it to see, while the store records every order. What we aim for is a known and stable gap, understood and explained, rather than an unexplained and shifting one.

We start every engagement by reconciling a month of orders from the store against the analytics report, line by line where needed, and writing down what explains each difference. That document becomes the foundation for every later decision.

The ones that describe the shopping path from first look to payment. In GA4 the standard ecommerce events include viewing an item, viewing a list, adding to cart, beginning checkout, adding shipping and payment details and purchasing. Each carries the product, price, quantity and currency, so that reports can show which items sell and where shoppers drop out.

We add the ones particular to your store: a size guide opened, a delivery estimate checked by pin code or ZIP code, a coupon tried and rejected, a payment method chosen and a payment that failed. These small events often show the real problem, such as a coupon field that sends people hunting for codes or a payment method that fails on certain banks.

Our GA4 setup and audit page describes the set-up, and the GA4 guide for small businesses is a useful introduction for those who have never configured it.

A tag manager holds all the tracking code in one place, so that changes do not require a developer to edit the theme and a broken tag can be removed in minutes. We set up the container with naming rules, a version record and a test routine, which keeps the store tidy and the load light. The Google Tag Manager setup page explains the method.

Browser-based tracking loses some data to privacy settings and blockers, and ad platforms are most affected because their reports depend on it. Server-side sending forwards purchase events from your own server, which can improve how completely and reliably orders reach the platforms. It adds cost and complexity, so we recommend it only where ad spend is large enough to justify it.

Consent matters here. We set up cookie banners and make tags respect the visitor's choice, to the specification your adviser gives us. Our page on server-side tracking and the Conversions API describes the work in more detail.

By reading the funnel as a series of questions. What share of visitors view a product, of those add to cart, of those begin checkout and of those pay? Splitting by device, by source and by new or returning shoppers shows where the biggest leak is, and it is usually much bigger on one phone type or one traffic source than on average.

Numbers say where, and not why, so we add evidence. Session recordings and heatmaps show what people tap, where they stall and what they ignore. Customer service messages and comments reveal the questions the page failed to answer. Our heatmaps and UX analytics page covers the tools, and we are careful to respect privacy by masking form fields.

The output is a short written finding, not a pile of charts: for example, mobile shoppers reach checkout but stall at the address form, or visitors from one campaign land on a page that does not match its promise. Each finding comes with a proposed fix and an estimate of effort.

In a short list of familiar places, and we check them in order. On product pages the usual causes are missing size or delivery information, weak images and reviews buried below the fold. In the cart they are unexpected costs and unclear totals. In checkout they are forced account creation, long forms, too few payment options and errors that do not explain what went wrong.

Speed and mobile layout cut across all of these. A page that loads slowly or has a button hidden under a banner will lose people no matter how good its copy is. For visitors from social platforms, the in-app browser can break features that work in a normal one, so we test there too.

We also look at trust: whether a first-time visitor can find who you are, how to contact you, how delivery and returns work and whether the payment page looks safe. Our conversion rate optimisation page sets out the general method, and this page applies it to stores.

Yes, but with realistic expectations. A test needs enough orders to separate a real difference from chance, and a store with a few hundred orders a month cannot detect small changes in a few weeks. For such stores we favour bigger, bolder tests, sequential before-and-after comparisons with care taken to account for seasonality, and fixes that are obviously right and need no test.

Where traffic supports it, we run proper tests: one change at a time, a written hypothesis, a stated metric, a minimum run time and a stopping rule decided in advance, so the result is not chosen to flatter an idea. We record each test, including the ones that lost, so that lessons accumulate and nobody repeats them in a year.

Our A/B testing page describes the method. We do not promise uplift, and a test that finds no difference is a useful answer, because it saves you from a costly change.

Fewer things than most dashboards do, and the right ones. Revenue and orders are the starting point, but a store also needs new versus returning customer revenue, average order value, conversion rate by device and source, return and cancellation rates and, where you can supply costs, contribution after advertising.

We build the dashboard in Looker Studio from your store and ad data, with one page for the owner that fits on a screen and a deeper set of pages for the marketing team. Each figure is defined in plain words on the page, so that nobody argues over what conversion rate means.

The key discipline is to show blended results next to platform results. If ad spend rose and total store revenue did not, the dashboard should make that visible. Our marketing dashboards page covers this work, and the monthly numbers feed the reports for performance marketing and automation on this industry's pages.

Several, and each is for your legal or compliance adviser to decide for your business. Analytics, remarketing and session recording all process personal data, and laws such as India's Digital Personal Data Protection Act and the state privacy laws in the United States affect how consent, notices and retention should work. Accessibility is another: in the United States, the Americans with Disabilities Act is applied to websites in many claims, and a store should be usable by keyboard and screen reader.

What we do is build sensibly. Consent tools that respect the visitor's choice, masked form fields in recordings, short retention settings, and layouts and forms that are readable and operable without a mouse. We do not claim that any of this makes a store compliant.

Our guides on the DPDP Act and your website and ADA website compliance are general starting points for conversations with your adviser.

With a tracking audit, which is a fixed, self-contained piece of work that you can take away. It reconciles orders, lists the broken or missing events, ranks the likely conversion problems and says what we would fix first. Some brands act on it themselves, and that is a fair outcome.

On proof, we have not published a case study of analytics or conversion work for a store, and we do not quote uplift figures. Related work is the website we built for Saptgiri Capital, a financial firm and not a store, where speed and a working enquiry form were central. It shows how we build, nothing more.

You can judge us on the method on this page and on the audit, which costs little compared with a month of decisions made on wrong figures. Questions are welcome through our contact page.