Why domain knowledge matters (and why I had to learn it the hard way)

E-commerce metrics were the last thing on my mind when I began preparing for data analyst interviews. I was familiar with SQL and could create an effective Power BI dashboard. Then, during a virtual interview, someone asked me, “Why are our profits down this quarter?” I had no idea how to begin. I could query the data. I simply couldn’t read the story it told.

That’s the gap I’ve been filling ever since—not more terminology, but the business logic underneath the numbers. In e-commerce, each metric on a dashboard represents a stage in a customer’s journey: how they discovered the store, why they purchased (or did not), and whether they returned. When I started thinking about it that way, the numbers became less abstract.

E-commerce is also one of the largest employers of data analysts, and I’ve noticed that the same few metrics appear repeatedly—in job descriptions, interview questions, across D2C brands, marketplaces, and even SaaS companies with a checkout flow. So I decided to write down what I’ve learned in the same way I wish someone had explained it to me.

Here’s what I’m going to talk about: the metrics that are important at each point of the customer journey—getting people to the site, getting them to buy, and getting them to return—and how analysts track them on a daily basis.

Acquisition metrics: bringing customers in the door

I began with these e-commerce metrics first, since it’s the most logical place to start—before anyone can purchase anything, they have to arrive at the location.

Customer Acquisition Cost (CAC)

CAC = Total marketing and sales spend ÷ Number of new customers acquired

The moment it clicked for me was: if CAC is more than what a customer ends up spending with you, you’re losing money on every new signup, no matter how great the product is. I also discovered that analysts rarely look at a composite CAC number—they split it down by channel (Instagram Ads vs. Google Search vs. affiliate) to evaluate which channel is genuinely efficient and which one is quietly wasting budget.

Return on Ad Spend (ROAS)

ROAS = Revenue generated ÷ Amount spent on advertising

This is the easier one: A ROAS of 4 means that every rupee you spend on ads generates four rupees in revenue. What shocked me is how often marketing teams utilize this data, not quarterly, but daily, to decide where to allocate expenditure. A channel heading down on ROAS is about to lose its budget, and calling out that pattern early is precisely the kind of work an analyst can own.

Conversion & value metrics: the checkout experience

Once I had the acquisition-side e-commerce metrics down, my next question was, “Okay, so people are coming—now what?”

Conversion Rate (CR)

CR = (Visitors who buy / Total visitors) × 100

Most e-commerce sites convert between 1 and 4%, which astonished me. It’s such a small number, which is why companies worry over A/B testing, tweaking a button color, a shipping message, or a checkout step, just to bump up their conversion rate by a fraction of a percent. I read that at scale, a 0.3% boost in CR can translate into crores of more revenue per year, which has changed the way I look at “small” UX adjustments.

Average Order Value (AOV)

AOV = Total Revenue / Total Number of Orders

That’s when I started to see trends in my shopping behavior; those “add ₹200 more for free shipping” and package offers are not random. These are deliberate nudges, as it’s typically cheaper to attract an existing shopper to add one more item than it is to gain a new customer.

Cart Abandonment Rate

Cart Abandonment Rate = (Carts Created − Completed Purchases) / Carts Created × 100

This statistic is in the 70s internationally, which honestly startled me the first time I read it; most “almost sales” never close. I’ve begun looking at the cart abandonment rate as a friction detector: unexpected shipping expenses at the final step, a forced account creation, and a payment failure. When this number declines, each percentage point often represents pure, regained revenue.

Quick gut-check: if your cart abandonment rate is over 70%, look at shipping-cost transparency and forced sign-ups first — they’re the two most common leaks.

Retention metrics: the key to profitability

Everyone talks about acquisition, but the more I studied these e-commerce metrics, the more I understood that the real profit is in retention.

Customer Lifetime Value (LTV / CLV)

is not simply the first order, but the overall net profit a company expects to make from a customer throughout the whole relationship. This is the measure that took me the longest to actually figure out. Because it’s not a historical sum, it’s a projection. What you’re doing is predicting how long a consumer stays with you and how much they spend—real prediction, not just doing some math.

The “Golden Ratio”: LTV : CAC

You can tell if the acquisition engine is healthy by comparing customer revenue to acquisition cost. 3:1 was the standard I kept reading everywhere, but going a little further, I realized that statistic actually comes from SaaS, where margins are strong and there’s no physical product to supply. E-commerce involves real costs of goods sold (COGS), including shipping, returns, and warehousing; therefore, its profit margins are inherently smaller. I’m currently treating the 3:1 ratio as a rough anchor rather than a strict rule, and I am verifying it against the gross margin before accepting it as valid.

Rule of thumb: don’t trust an LTV:CAC ratio until you’ve checked it against gross margin, not revenue.

To make sure I actually understood this (and not just the formula), I built a small hypothetical example for myself:

What stood out while building this model is that the gross margin is the key factor that makes the entire margin meaningful. If I’d compared LTV on revenue alone—skipping COGS entirely—I would have overstated how healthy the unit economics actually are. That’s a mistake I want to avoid making in an interview.

The analyst’s toolkit: how I’m starting to track these

Understanding the definitions was my initial step. The harder part has been figuring out how these e-commerce metrics actually show up in day-to-day analyst work.

SQL

This is the kind of query I’ve been practicing on sample transactions data from my dataset library, to compute AOV by month:

SELECT
  DATE_TRUNC('month', order_date) AS month,
  SUM(order_revenue) / COUNT(DISTINCT order_id) AS aov
FROM orders
GROUP BY 1
ORDER BY 1;

Dashboards (Tableau / Power BI)

I’m still learning these tools, but the concept makes sense: KPIs like conversion rate and CAC are turned into automatic visual reports that stakeholders can access on a regular basis without having to contact an analyst.

A/B Testing

This is the portion I’m most excited to delve more into. When a product team needs to know whether a new checkout button color increased conversion rates or if the change is merely noise, it’s a statistical inquiry. Before implementing a modification, analysts conduct t-tests (or a chi-square test for a rate like CR) to see whether the increase is statistically significant. I’m now testing this logic on my practice datasets.

Frequently asked questions

Conclusion

Writing these e-commerce metrics down helped me see them as a connected tale rather than a list of statistics to memorize. When the conversion rate is close to zero, high traffic is meaningless. A high conversion rate is meaningless if CAC consumes every rupee of profit. That’s the kind of thinking I’m trying to build—going from “I can run the query” to “I can tell you what the business should do next.”

“Try it yourself — for data aspirants, thru my links: the Brazilian E-Commerce Public Dataset (Olist) and my own dataset library.”

About the author
I'm Priyanka, a self-taught data learner and explorer. I share what I learn on BloomInData.