Even if you master Python to a professional level and your Excel spreadsheets look like works of art, you won’t get far in the world of data analysis without understanding analytical terminology. I repeat this to my students all the time: technology is just a tool. The true value of an analyst comes from the ability to navigate metrics fluently and understand what we are measuring and why.
In today’s business world, terms like CAC, Lead Time, or Contribution Margin play a crucial role. Of course, much depends on the industry you work in, but there are fundamentals that you simply have to “get” if you want to be a partner for the business, rather than just someone who clicks through reports. I decided to gather fifteen of the most important Key Performance Indicators (KPIs) and divide them into three difficulty levels, corresponding to three different industries: from accessible e-commerce, through slightly more demanding logistics, to the expert level—finance.
Grab a coffee, because we have a lot of concrete knowledge ahead of us. We will go through not only dry definitions but, above all, business examples and the most common mistakes analysts make when calculating them.
Level 1: E-commerce – The Foundations Where Everything Begins
The e-commerce industry is the perfect place to learn analytics because almost everything here is measurable. Data flows in a wide stream, and IT systems often deliver basic calculations themselves. However, just because a system calculates something doesn’t mean it does it correctly or that we understand the result.
1. CAC – Customer Acquisition Cost
This is the absolute baseline. CAC simply tells us how much money we have to spend to acquire one new customer. Sounds simple? In theory, yes: you divide total marketing costs by the number of customers acquired. However, the devil is in the details.
The most common mistake I encounter is mixing acquisition channels. If you spend significantly more on ads in one channel than another but calculate CAC globally, you lose sight of the effectiveness of individual actions. Another issue is ignoring operational costs related to setting up the purchase path. Also, remember the scaling phenomenon—when you increase advertising budgets, the cost of acquisition often rises because you start reaching “colder” audiences who are harder to convince. If an analyst fails to catch this, the business may start burning through its budget at an alarming rate.
2. LTV – Customer Lifetime Value
LTV determines how much money a given customer will leave with us during their entire relationship with our brand. This is the metric that gives meaning to CAC calculations. The best business is a repeatable business. Many people wrongly assume that LTV only counts in subscription models like Netflix or my KajoData Space. Nothing could be further from the truth.
LTV can and should be calculated even for fast-moving consumer goods. Take the example of Coca-Cola—if you buy it regularly every week, the brand can precisely calculate your value over the years. For an analyst, the key is connecting LTV with CAC. The LTV to CAC ratio shows whether the business model is profitable at all. Imagine two marketing channels: Channel A has a higher CAC, but customers from it buy four times more often than those from Channel B. Without knowing the LTV, you might mistakenly recommend cutting the budget for Channel A, when that is exactly where we acquire the most valuable, loyal customers.
3. AOV – Average Order Value
We want to know how much the average customer leaves in their basket during a single transaction. It’s a simple division of revenue by the number of orders, but analysts often forget what distorts this average. The biggest enemy of a reliable AOV is periodic promotions.
If your company regularly organizes events like Black Friday, back-to-school, or New Year’s sales, you must calculate AOV broken down into “normal” and “promotional” periods. The real challenge begins when the price distribution in your store is very wide—from products for $5 to those for $500. In such cases, AOV helps decide which products are worth getting rid of (because they generate maintenance costs without building basket value) and which are worth scaling.
4. Cohort Retention Rate
This is a metric I love to use in KajoData Space. It involves grouping customers by the time they started using the service (e.g., the January 2025 cohort). This allows me to check what percentage of people from that group are still with me after a month, three months, or a year.
The biggest mistake is confusing retention with purchase frequency. Someone might buy a lot of products from you in one month because they are excited about a new brand, and then disappear forever. This means you didn’t offer them long-term value. Retention decides whether you have to spend atomic amounts on constantly acquiring new people or if you can build a business on the loyalty of existing ones. In SaaS (Software as a Service), a low retention rate is an alarm signal—something is wrong with the product if customers leave despite high switching costs.
5. Customer Payback Period
A metric often confused with LTV, but in my opinion, much more interesting from an operational perspective. It tells us how long it takes before we “break even” on a given customer. In e-commerce, where we often want to scale quickly with expensive ads, we forget about operational costs and margins.
LTV tells us how much we will earn in total, but the Payback Period tells us how long we must keep a customer just to cover the cost of their acquisition and service. If a customer has to stay with you for six months to stop generating a loss, and your average retention time is five months, you have a massive problem. A business must reach break-even relatively quickly, especially if you don’t have the scale of giants like Netflix.
Level 2: Logistics – Mathematical Precision and Brutal Reality
In logistics, there is no room for interpretative errors. Here, everything is based on dates, inventory levels, and the physical flow of goods. Small errors in calculations at a large scale can have colossal financial consequences.
6. Return Rate
Although associated with e-commerce, from an analytical perspective, this is a logistical challenge. If you offer a return guarantee, you must precisely calculate how much merchandise is coming back to you. Revenue in the bank account might be pleasing after a successful campaign, but the Return Rate is a boomerang that can come back after two weeks and “cut your business head off.”
It is important to break down returns by individual products and procedures. If a new clothing line has a 30% return rate because the sizing is incorrectly described, your great revenue is just an illusion of profitability. An analyst must be the voice of reason here, cooling the enthusiasm of the sales department.
7. Lead Time
This is the difference between the date an order is placed and the date it is delivered to the customer. In an era where the fight is for the package to be with us “by tomorrow,” this is a key KPI. The challenge is that we often do not have full influence over Lead Time when using external courier companies.
An analyst’s most common mistake is looking at Lead Time too generally. Let’s take an example from my experience—an e-commerce operating in Europe. You look at the averages and see that Greece has terrible results. Does this mean the country manager is failing? Not necessarily. If you don’t break down the data to a granular level, you won’t notice that Greece consists of thousands of islands where transport takes place on small boats that are not as predictable as trucks. Only such an analysis allows for sensible business decisions.
8. Order Cycle Time (OCT)
Similar to Lead Time, but focused on what happens inside your warehouse. We measure the time from clicking “buy” to the moment the package is ready for the courier. This is the segment over which we have full control.
OCT changes drastically during peak periods like Black Friday. The warehouse swells, bottlenecks appear, and throughput drops. An analyst must be able to isolate these processes to indicate exactly where we are losing time—whether it’s during packing or perhaps during order picking.
9. Fill Rate
If you run a warehouse, you want to fulfill 100% of orders. If you don’t, backorders are created—orders waiting for missing stock. The key to understanding Fill Rate is the concept of SKUs (Stock Keeping Units).
A global Fill Rate of 98% might sound great, but what if those missing 2% are your bestsellers with the highest margin? An analyst must analyze Fill Rate broken down by top products. You can have a warehouse full of things nobody wants and a great Fill Rate for them, but lose a fortune on the lack of key items.
10. Cost per Shipment
This is a complex metric. You can calculate it simply—per courier—but you can also go deeper, including the costs of the warehouse, people, and packaging. Interestingly, analytics often shows that the biggest cost in a shipment isn’t transport, but… air.
Unoptimally packed boxes mean you are paying for transporting empty space. Cost per Shipment provides information about scalability. If sales grow but shipping costs grow even faster, your business will eventually stop working under the operational weight. Always ask yourself: what exactly are we including in this cost? Are we forgetting about boxes, tapes, or hall maintenance costs?
Level 3: Finance – Where Costs Become Elusive
In finance, we enter the highest level of difficulty. While in logistics you see empty shelves in a warehouse, in finance, costs can slip away in a way that is almost invisible to the untrained eye.
11. COGS – Cost of Goods Sold
This is the basis for calculating margins. It’s about how much we have to take out of our own pocket to be able to sell something. In the case of digital products, like my courses on KajoData, COGS can be elusive if we don’t count labor time.
I strongly encourage analysts and solopreneurs to value their own hourly rate and include it in the cost of producing a good. If you spend 100 hours creating a course, that has a specific financial value. Only by including this “invisible” cost can you reliably calculate the margin and predict the budget for the future.
12. Contribution Margin
This is a metric that allows us to evaluate our promotional capabilities. It involves dividing costs into fixed and variable. Contribution Margin is sales revenue minus variable costs (those that grow with the scale of sales).
Why is this important? Because every promotion eats your margin. The lower the Contribution Margin, the less money you have left to cover fixed costs (rent, servers, salaries). An analyst must know how far one can go with discounts without starting to pay extra for the business.
13. OPEX – Operating Expenses
Anyone who has worked in IT knows this term. OPEX is the cost of running the business—salaries, software, office, legal services. It is the money needed to keep the machine in motion.
You can have a brilliant product with low COGS and great marketing, but if your OPEX is “lethal” (e.g., due to an over-expanded office structure or unnecessary tool subscriptions), the business will fail. OPEX doesn’t have a single formula, but an analyst must monitor it so that the machine in the back doesn’t become too heavy for the business idea itself.
14. Net Profit Margin
My favorite KPI. This is the ultimate truth test for a business. It shows what percentage of revenue constitutes real net profit. It is the ratio of profit to revenue.
We can brag about million-dollar revenues and great ROI from ads, but if after paying COGS, OPEX, and variable costs our Net Profit Margin shrinks or stagnates, then scaling makes no sense. Not everyone is Uber or OpenAI, able to afford years of losses. In smaller companies and in the solopreneur model, NPM is the indicator that determines your “to be or not to be.”
15. Fixed Variable Cost Ratio
Finally, something for true analytical connoisseurs. The FV Ratio doesn’t require complicated formulas, but it requires a deep understanding of the business. We consider the proportion between what we always pay and what depends on the sales volume.
What grows faster when the business goes up? If fixed costs grow too quickly, you fall into the volume pressure trap—you have to sell more and more just to stay afloat. Conversely, a well-optimized business is one where fixed costs stay in place, and only variable costs directly related to revenue grow. An analyst must work closely with the business to correctly classify costs—for example, is marketing a fixed cost (team salaries) or a variable one (advertising budget)? The answer depends on the company model.
Summary: More Than Just Excel
I hope this overview made you realize one thing: a data analyst’s job is not just about knowing the XLOOKUP function in Excel or joining tables in Pandas. Business doesn’t need someone who just “makes reports.” Business needs someone who understands reality through data.
Understanding that a customer has a lifecycle (LTV), a cost (CAC), and that revenue is just the beginning of the conversation about profitability is the absolute foundation of your professional development. If you can combine technical skills with such business understanding, you will become a partner for your company that cannot be easily replaced.
Thanks for joining me in exploring these 15 indicators. If you found this knowledge valuable, please share this article on your social media—on LinkedIn, Facebook, or Twitter. Let this knowledge—that an analyst is someone who “figures out reality”—go out into the world!
The article was written by Kajo Rudziński – analytical data architect, recognized expert in data analysis, creator of KajoData and polish community for analysts KajoDataSpace.
That’s all on this topic. Analyze in peace!
Did you like this article 🙂?
Share it on Social Media 📱
>>> You can share it on LinkedIn and show that you learn something new every day.
>>> You can throw it on Facebook – and perhaps help a friend of yours who is looking for this.
>>> And remember to bookmark this page, you never know if it won’t come handy in in the future.
You prefer to watch 📺 – no problem
>>> Subscribe and watch my English channel on YouTube.
Prefer to read in Polish? No problem!
Other interesting articles:
