By Stormly  in  Knowledge

Published: Nov 16, 2026

Product Analytics Software for eCommerce: What You Actually Need, What to Skip, and How to Choose

Most eCommerce teams start looking for product analytics software after the same realization: Shopify tells them total revenue, GA4 tells them traffic sources, and neither one answers the question they actually need answered. Which products are building repeat customers? Which products are killing the checkout? Which pairs sell together?

The category is genuinely confusing because “product analytics” was invented by and for SaaS companies. Mixpanel, Amplitude, Heap, and most tools in this space were designed to track feature adoption, onboarding funnels, and user activation flows for software products. The features eCommerce stores actually need look different, and so does the evaluation checklist.

Product analytics software for eCommerce is a category of tools that answer product-level questions: which SKUs contribute the most revenue, where shoppers abandon by product not just by funnel step, which products sell together, and how cohorts of first-time buyers behave based on their first purchase. The right tool treats your order history and product catalog as first-class data, not a custom event taxonomy you configure yourself.

At a glance: - Most product analytics tools were built for SaaS; eCommerce needs a different feature checklist - Must-haves: SKU-level revenue contribution, cart abandonment by product, cross-sell pairs, cohort retention by first purchase - Nice-to-have but skippable for most stores: in-app guides, session replay, feature flags, DAU/MAU session counters - Shopify reports and GA4 answer different questions; they do not cover the product-level layer - Best evaluation method: bring 3-5 specific store questions and test them against your own data in a trial

What is product analytics software?

Product analytics software tracks and analyzes how customers interact with products. For a software company, that means feature adoption, onboarding completion, and in-app behaviors. For an eCommerce store, it means which products convert, which ones drive repeat purchases, which products are bought together, and which products cause checkout abandonment.

The core difference from web analytics is the unit of analysis. Web analytics tools count sessions, pages, and events. Product analytics counts orders, SKUs, and customer journeys across purchases. A session ends when the browser closes; a customer relationship spans years and many orders. The right product analytics tool is built around that second frame.

What product analytics is and what it can actually do for an online store covers this distinction in detail, including why borrowing SaaS analytics frameworks for a Shopify store creates as many blind spots as it solves.

What features does an eCommerce team actually need?

Here is where the SaaS vs eCommerce split matters most. SaaS-native product analytics tools ask you to define an event taxonomy: label every user action, map your data model, build your funnels from scratch. That is powerful for software products. For a store running on Shopify or Adobe Commerce, your order history already knows what matters. The question is whether the tool can read it.

The five features that actually move decisions for eCommerce teams:

Revenue contribution by SKU and category. Not total revenue. Not revenue by channel. Revenue by product, filtered by category, brand, attribute (size, color), and time period. This tells you which 20% of the catalog is generating 80% of the revenue, and which products are carrying the store vs riding along.

Cart abandonment by product. A site-wide checkout rate of 60% is a weighted average. Some products complete at 74%; others complete at 29%. The products dragging down the average are invisible in a site-level metric. Cart abandonment by product shows add-to-cart rate, abandonment rate, lost revenue, and recovery potential per SKU or category.

Cross-sell pairs and basket analysis. Which products are bought together, and by how much more than baseline? The useful metric is lift: how much more often this pair occurs than you would expect if purchases were random. Attach rate and co-order volume round out the picture. These numbers feed bundle offers, frequently-bought-together widgets, and email sequences.

Cohort retention by first purchase. This is the question most stores cannot answer: of the customers who first bought Product A vs Product B, who came back within 90 days? First-purchase product is usually the strongest predictor of long-term customer value, and it is invisible in tools that only track aggregate retention.

Plain-language questions. Most teams do not have a data analyst available for every question. A tool that lets a merchant type “which product category had the most recoverable cart abandonment revenue last week?” and get a usable answer has a real productivity advantage over one where every question requires building a custom report from scratch.

The table below shows what to look for in a trial for each of these:

Need Why it matters for eCommerce What to test in a trial
Revenue by SKU, category, attribute (illustrative: top 20 SKUs drive 78% of revenue) Hides the catalog distribution most aggregate reports obscure Filter to last 90 days; sort by revenue and by unit count separately
Cart abandonment by product Site-level conversion rate is a weighted average that masks the problem products Filter by SKU; look for the gap between add-to-cart rate and checkout completion
Cross-sell pairs (lift + attach rate) Identifies bundle and upsell sequences that exist in the data Sort by lift, not just co-occurrence volume
Cohort retention by first purchase Predicts which products build loyal customers vs one-time buyers Segment a retention cohort by first-purchase product
Plain-language or MCP questions Reduces dependency on data analysts for routine questions Ask a specific product question in natural language and check the answer against a known metric

Stormly’s Revenue Contribution Analysis, Cart Abandonment report, and Cross-Selling Analysis report address each of these using your store data as it comes in, without requiring you to design an event taxonomy first. The MCP layer lets you ask questions in plain language from Claude, ChatGPT, or any compatible MCP client.

What eCommerce stores can skip

Most product analytics platforms include features that are genuinely useful for SaaS products but add noise for stores:

  • In-app guides and tooltips: useful for software onboarding, not relevant for checkout flows
  • Session replay: helpful for UX debugging but not for catalog-level decisions
  • Feature-flag management: a SaaS deployment concept, not relevant to eCommerce product strategy
  • DAU/MAU session ratios: the right engagement metric for software products; for a store, repeat purchase rate by cohort is more meaningful

If a tool’s main use case examples are “track feature adoption” and “measure onboarding completion,” it is designed for a software product.

[Run the checklist on your own data. Stormly’s free trial connects to your Shopify or Adobe Commerce store without requiring event taxonomy setup.]

How does product analytics software compare to GA4 and Shopify reports?

The most common confusion in this category is the assumption that GA4 or Shopify Analytics already does what product analytics does. They don’t, but they each do other things well.

Tool What it answers well What it does not answer
Shopify built-in reports Total revenue, orders by channel, gross margin by product, store-level conversion rate Which product has 29% checkout completion vs 74%; which first-purchase SKU predicts a second order
Google Analytics 4 Traffic sources, page engagement, goal completions, session attribution Product-level cohort retention, cart abandonment by SKU, cross-sell lift, revenue by attribute
Product analytics (eCommerce-native) SKU revenue contribution, cart abandonment by product, frequently bought together, cohort retention by first purchase, anomaly detection Multi-touch ad attribution, session replay, A/B testing (unless separately bundled)

None of these is better in absolute terms. They answer different questions. The gap is specifically the product-level layer: the analysis that ties individual SKUs to customer lifetime value, repeat purchase rate, and checkout abandonment by product. That layer is not covered by session tools or marketing attribution tools.

Where your Shopify checkout actually leaks by product shows a worked example of what that product-level analysis looks like in practice, using cart abandonment data to surface which specific SKUs have the worst checkout completion and what the abandoned revenue figure is.

How do you evaluate product analytics software in a two-week trial?

The worst evaluation method is reading feature lists and screenshots. The best method is bringing 3-5 questions your team has been trying to answer, connecting your store data, and seeing whether those questions are answerable in under 20 minutes.

Here are five questions that reliably reveal whether a tool is genuinely built for eCommerce:

  1. Which 10 SKUs contributed the most revenue in the last 90 days, and what is each one’s repeat purchase rate?
  2. Which products have the highest add-to-cart rate but the worst checkout completion?
  3. Which product pairs sell together most often, and what is the lift above random baseline?
  4. Of customers who first bought Product A vs Product B, who had a higher 90-day return rate?
  5. Which product categories have the most recoverable cart abandonment revenue right now?

If the tool requires you to build a custom event taxonomy and wait several weeks before those questions are answerable, the data model was not designed for your use case. If the tool answers all five in a trial against your real store data, that is a strong signal the architecture is right.

For teams evaluating multiple tools side by side, product analytics tools compared for eCommerce gives a capability matrix organized by those eCommerce-specific jobs. For a broader overview of the field before committing to a trial, the best product analytics tools for eCommerce teams covers the main options ranked by what store operators actually need.

What about plain-language questions and MCP?

A meaningful shift in this category is the ability to ask store questions in plain language, either through a chat interface or through the Model Context Protocol (MCP), which connects your store data to tools like Claude or ChatGPT.

The key distinction is not whether a tool has a “chat” button. It is whether the underlying data is modeled well enough that the question you type maps to an accurate answer. A raw API connection to your Shopify Admin will return numbers but won’t understand what “cart abandonment by product” or “retention by first purchase” means without a layer that models those concepts correctly.

Stormly’s MCP server routes plain-language questions to its matching reports or custom SQL, using a data model built around commerce concepts. Example questions from the docs: “Which SKUs grew revenue fastest last week compared with the previous week?” and “Which product categories have the most recoverable cart abandonment revenue?” Those questions work because the data layer underneath already understands products, carts, orders, and cohorts.

For eCommerce teams without a dedicated analyst, that difference matters in practice.


The evaluation checklist for product analytics software is simpler than it looks once you remove the SaaS-native features you don’t need and focus on the five jobs that actually drive catalog and retention decisions. SKU-level revenue contribution, cart abandonment by product, cross-sell pairs ranked by lift, cohort retention by first purchase, and plain-language access to those numbers: that is the core of what a product-level analytics layer does for an online store.

Stormly is built from the ground up for those five use cases. It connects to Shopify, Adobe Commerce, Segment, Google Tag Manager, and other commerce data sources.

Run the checklist on your own data. Start your free trial.

Ready to get real insights?

Connect your store and let Stormly's AI find the trends and anomalies that matter.

No credit card required