Product Hierarchy: How to Link UX Solutions to Business Metrics

The designer redesigned the onboarding. Users began to pass it faster. The designer is pleased. The product looks at revenue - it hasn't grown. "Why did you even do that?"

This is the classic gap between UX metrics and business metrics. The designer optimized one thing, the business expected another. It was no one’s fault, there was no common language.

The hierarchy of metrics is that common language. The system that shows that business wants X, X is made up of Y and Z, Y is directly dependent on what the designer does.

## Why do we need hierarchy at all

Metrics without hierarchy are a set of numbers. Everyone looks at their own: marketing on CAC, product on retention, designer on task completion rate. The numbers do not contradict each other, but do not add up to the picture.

Hierarchy answers the question, “Why is this metric important?” Not just “retention is up 3%” but “retention is up 3%, so in six months’ time, revenue will grow by about X, because that’s the link.”.

Three things that hierarchy gives:

Prioritization. When it’s clear what’s going on, it’s easy to decide what to do first. Not by sensation, but by chain.

** Reasoning for decisions.** “I want to redesign the registration form” is weak. The sign-up form is a bottleneck in the funnel that cuts conversion rate, and CR directly affects revenue.

** Conflict detection. ** Sometimes optimizing one metric hurts another. Hierarchy helps us see this beforehand.

## Three metric levels

Any product can be described in three levels of metrics. They are called differently, but the essence is the same.

### Level 1: Business metrics

This is what the CFO and CEO believe. Money, growth, business sustainability.

• Revenue / MRR / ARR – revenue, monthly and annual
• Profit margin - margin
• **LTV (Lifetime Value) - How much money does one user make over time
• **CAC (Customer Acquisition Cost) - How much does it cost to attract one user
• LTV/CAC ratio - the main indicator of business health (the norm is 3:1 and above)
• Churn rate – Percentage of users who leave

At this level, the designer usually does not work directly. This is where you need to know what is considered a success.

### Level 2: Product metrics

This is what the product team measures. Indicators of user behavior in the product.

• DAU/MAU – daily and monthly active users
• *Retention Rate – What percentage of users return (D1, D7, D30)
• Activation Rate – Have users reached the point of value
• Engagement - Depth of use of the product
• **NPS (Net Promoter Score)
• Feature adoption - whether specific features are used

Here, the designer already influences directly - through UX, information architecture, onboarding.

### Level 3: UX metrics

That's what a designer measures. Quality indicators of a specific interaction.

• Task success rate - Percentage of users who have successfully completed a task
• **Time on task ** How long does it take to complete
• Error Rate - How often do users make mistakes
• **SUS (System Usability Scale) - subjective assessment of convenience
• **HEART Metrics ** Happiness, Engagement, Adoption, Retention, Task Success
• Conversion rate in the funnel - the percentage of transition between steps

These are metrics of specific screens, flow, components.

## How to Link Levels: Tree Metrics

The hierarchy works like a tree. The top level is business metrics. It is decomposed into product metrics. These are decomposed into UX metrics.

Here is an example for a SaaS product:

Revenue
── New Paying Users
← Traffic to the website
Conversion rate (landing → registration)
Task completion rate on the registration form
← Error Rate on the Steps of the Form
Activation Rate (Registration → First Valuable Action)
Time to Value (how quickly the user gets to value)
Onboarding completion rate
Retention of existing users
─ D30 Retention
← DAU/MAU ratio
Feature adoption rate of key features
Expansion revenue (upgrades)
─ Use of premium feature
Frequency of use

This isn’t the only possible type of tree – it will be different for different products and business models. But the principle is the same: each top-level metric is explained through lower-level metrics.

### How to read a tree

From top to bottom: What affects revenue? New users and retention. What affects the Conversion rate? → Task completion on the registration form and error rate.

Bottom up: “We have improved the error rate on the registration form. What does this mean for business? “Higher conversion rate → more new payers → higher revenue.”.

The second way is how the designer justifies his work to the business.

## How to build a tree for your product

### Step 1: Set your main business metric

It's North Star Metric or just the main KPI that matters to the business right now. Not three metrics and not "well, we care about everything" - one.

Examples: Monthly Recurring Revenue, number of active users, volume of transactions.

### Step 2: Break it down into components

Any metric can be mathematically decomposed. Revenue = Number of users × ARPU. DAU = New Users + Returnees - Left. Retention = 1 − Churn.

These are not hypotheses; these are mathematical identities. Use them as a frame.

### Step 3: Find a product driver for each component

“What does the product have to do with this component?” Retention usually depends on engagement with key features, on the quality of onboarding, on how quickly the user achieves value.

### Step 4: For each product driver, find a UX metric

"How do you measure that in the interface?" Feature adoption rate, time spent in feature Onboarding completion rate, time to completion, drop-off points.

### Step 5: Check the links

The relationship between levels must be logically justified. “If the onboarding task completion rate rises, the activation rate should rise too,” is a testable hypothesis.

## Common mistakes in building a hierarchy

### Mistake 1: Too many metrics on each level

If you have 40 metrics at the UX level, it's not a hierarchy, it's a data table. Each level should have 3-7 key metrics.

### Mistake 2: Connections “by sensation”

"We think it affects that" is not good. Relationships must be either mathematically inferred or confirmed by data. Hypotheses should be explicitly labeled as hypotheses.

### Mistake 3: A static tree

The product changes, the focus of the business changes – and the metric tree has to change. A quarterly audit is a minimum.

### Mistake 4: Focus on what is easy to measure

Clicks, views, time on the site is easy to calculate. But they often don't correlate with what's important to a business. Build a hierarchy from the business goal down, not from the data up.

### Mistake 5: Confusion of Metrics and Goals

Metrics are dimensions. The goal is the desired value of the metric. "Retention" is a metric. "Retention D30 ≥ 40%" is the target. Don't mix them up in a tree.

## Designer and metrics: three models of interaction

In different teams, the designer works with metrics differently.

### Model 1: Designer gets metrics on top

The product sets a goal with a metric: “We need to raise the conversion rate on onboarding from 60% to 75%.” The designer solves the problem.

Pluses: A clear task, clear criteria for success. *Minuses: * The designer doesn't understand the context, can optimize the wrong way.

### Model 2: Designer participates in metric selection

Together, the team decides what to consider a success for a particular project. The designer offers UX metrics that are related to product goals.

Pluses: The designer understands the whole task, the metrics are meaningful. Minuses: requires team maturity and trust.

### Model 3: Designer builds and offers metrics

The designer comes to the product with the finished chain: “I propose to change the form, this should raise the task completion rate, which is associated with the activation rate, which directly affects revenue.”.

**Pluses: ** Maximum impact and responsibility. Minuses: You need a deep understanding of business and data.

The third model is the senior/lead designer level. That’s why you need a hierarchy of metrics.

## How different types of products build a hierarchy

### E-commerce

GMV (Gross Merchandise Value)
Number of orders
← Conversion Rate (Session → Order)
← ← Add to cart rate
← Cart abandonment rate
← Checkout completion rate
← Repeated orders
Retention rate of buyers
Average check
Average basket size
Upsell/cross-sell rate

The designer here works primarily with the funnel - from the product card to the order confirmation. Key metrics: add to cart rate, checkout completion rate, cart abandonment rate by steps.

### SaaS (subscription)

MRR (Monthly Recurring Revenue)
New subscriptions
Trial-to-paid conversion
Activation rate (trial → aha moment)
Engagement during the trial
← Number of Trial Registrations
Retention of subscriptions
─ Churn rate
Feature adoption (do key features use)
Engagement trend (grows or falls)
Expansion revenue (upgrades)

A key design challenge in SaaS is to accelerate the user’s journey to the “aha moment” and ensure constant engagement with the features they pay for.

### Marketplace (bilateral)

Marketplace is more complex – there are two sides and metrics for each.

GMV
Demand side (buyers/customers)
← Conversion Rate (Search → Order)
← Retention of Buyers
The average check
в ─ Party of offer (sellers/executors)
Supply Quality (Ratings, Reviews)
Retention of suppliers
Time to fill (time to fill)

A marketplace designer works with two different users with different tasks and often with different interfaces.

## When Metrics Lie: Traps for a Designer

The hierarchy of metrics helps not only to see connections, but also to notice when data is misleading.

### The “good metric, bad product” trap

Task completion rate of 95% on form is excellent. But users filled it out for 12 minutes and hated every second. SUS score 42/100.

Metrics do not contradict, but the picture is different. You need to look at several related metrics together.

### The Trap of “Correlation Instead of Causality”

Implemented a new dashboard - retention increased. I think there's a connection. But at the same time launched an email campaign and improved the mobile application. What exactly affected?

Without controlled experiments (A/B tests), it is impossible to establish a causal relationship. Correlation is a hypothesis, not a fact.

### The Vanilla Metrics Trap

Some metrics look good by definition. "We've been liked 10,000 times," so what? The number of likes alone does not explain whether a product helps users achieve their goals.

Vanilla metrics are the ones that are difficult to associate with business outcomes. They don't fit into the hierarchy.

### The Trap of Local Optimum Optimization

The conversion rate increased from 40% to 55%. Victory? Not if the activation rate fell from 70% to 50% because non-target users went into the product.

Optimizing an individual metric without understanding the system can degrade the outcome.

## Metrics and Design Solutions: How to Build Argument

When a designer comes up with a proposal for a product or business, the structure of the argument must follow a hierarchy of metrics.

Weak argument: I want to remake the payment page because it looks outdated.

Strong Argument: We have a 38% drop-off in the payment step. This reduces the checkout completion rate to 62%, which cuts the entire funnel conversion rate. According to our calculations, raising checkout completion to 75% will give +18% to the number of orders per month. I suggest reworking the payment step, starting with simplifying the form and adding a security indicator.”.

The difference is that the second argument goes down the tree metric from top to bottom. The business effect is clear, the proposal is specific.

## How to Start Right Now

If you do not already have a hierarchy of metrics, you can start in one working day.

Morning: Collect data. Take any analytical dashboard and write down all the metrics that are there. Don't judge, just collect.

**Dinner: ** Talk to a product or CEO. One question is, “What metric is the most important to us right now?” This is your upper level.

**Dinner: Build the first draft of the tree. Not perfect - just decompose: what affects the main metric, what affects it.

**Take a walk through the tree with the team. Where connections are not obvious, ask a question. Where something is missing, add it.

A black tree is better than no tree. It can be clarified as you work.

## The hierarchy of metrics is ready if..

• There is one key business metric at the top
• Each level has no more than 5-7 metrics
• Relationships between levels explained (mathematically or through data)
• For each UX metric, it is clear how to measure it
• A designer can explain any decision through a chain of metrics
• The tree is updated once a quarter or when the focus of the business changes
• Linkage hypotheses are clearly labeled as hypotheses
• The team is aligned by hierarchy (not only the designer knows it)

## What next

The hierarchy of metrics is the foundation. Everything else is built on it: the choice of North Star Metric, building a dashboard, A/B tests, reasoning design decisions.

The next step is to learn how to choose one main metric from the whole tree and not lose sight of the others. This is the North Star Metric theme, a separate article in this series.

## How metrics change when scaling a product

The hierarchy of metrics is not constant. As the product and business grow, the focus shifts – and the metric tree should change with it.

### Stage: Product-market fit (0→1)

At this stage, the question is, does the product work at all? Does it create value for users?

Key stage metrics:

• Retention of the first users (Is there anyone coming back?)
• NPS (Do they recommend the product?)
• Quality feedback

At this stage, Revenue is a poor master metric. Even with a bad UX, you can sell your first customers on enthusiasm. This does not prove that the product works.

UX metrics at this stage: Task success rate on key scenarios and time to first valuable result.

### Stage: Growth (1→N)

The product works, you need to scale.

** Focus shifts to:**

• Acquisition efficiency (CAC, conversion funnel)
• Retention at scale (Do we retain a diverse audience?)
• Activation (Do new users see value?)

UX metrics focus on attraction funnel and onboarding – there are the biggest losses when scaling.

### Stage: Maturity

The product is mature, growth is slowing down.

** Focus on:**

• Expansion revenue (more from existing users)
• Efficiency (reduced CAC, reduced maintenance cost)
• Defense (holding in front of competitors)

UX metrics are shifting towards engagement and depth-of-use: whether users are using the full potential of the product rather than just basic features.

## Metrics and Prioritizing Design Tasks

The hierarchy of metrics is not just a theory. It is a working tool for making decisions about what to do.

### How to Use Hierarchy to Prioritize Backlog

You have 15 backlog tasks. How do you choose where to start?

Step 1. For each task, determine which metric it should improve?

Step 2. Find this metric in the tree. What level is it at?

Step 3. What is the potential for improvement? If that metric goes up 10 percent, what happens up the tree?

** Step 4.** Evaluate the effort. Big influence + little effort = start from here.

### Example of prioritization

Task backlog:

• A: Remove the main page (affects: first impression, no clear metric)
• B: Simplify the application form (affects: conversion rate → revenue)
• C: Add a dark topic (affects: satisfaction, indirectly retention)
• D: Correct errors in the mobile payment form (affects: checkout completion → revenue)
• E: Improve onboarding step 3 (affects: activation rate → retention → revenue)

Prioritization through hierarchy of metrics:

1. D (Critical Payment Mistake, Direct Impact on Revenue)
2. B (conversion directly to revenue)
3. E (activation rate → retention → long-term revenue)
4. C (retention through satisfaction, indirect influence)
5. A (the impact is immeasurable without a specific metric)

It’s not a perfect algorithm — but it’s much better than feeling prioritized.

## How to talk about metrics to different audiences n

The designer works with different people: CEO, product, developer, other designers. Everyone has their own language.

### For the CEO

Speak in terms of money and risk. Not “retention grew”, but “retention grew by 3%, this is broadcast in +X rubles of revenue in 6 months according to our calculations”.

The CEO thinks, "Is it important?" How much is it? When will it pay off

### For the product

Talk about product metrics and hypotheses. “Activation rate of 38% is below the benchmark of 55%.” Assume that the reason for step 3 onboarding – there is 42% drop-off. We propose a new flow test.”.

The product thinks, "What are we doing, why is this, how do we measure the outcome?"

### For the developer

Talk about specific tasks and their priority. The error rate on the payment form is 12%, which is critical. Users cannot enter CVV on mobile. This blocks conversion.”.

The developer thinks, "What exactly needs fixing and how urgent is it?"

### For other designers

Talk about design decisions and their rationale. “We removed the telephone field from the uniform. The Task completion rate rose from 67% to 84%. The specific problem was that users did not understand why the phone was needed when registering online.”.

Another designer thinks, "Why is this particular solution, what are the alternatives, what can I apply to this?"

## AI and Metric Hierarchy: How Claude Helps Build a Tree

Building a tree metric from scratch is a task for several hours: you need to understand the business model, write metrics, find connections. AI reduces this to 20-30 minutes.

Here's how it works in practice.

### Prompt: Build the First Tree Metric

I am a product designer in [product type: SaaS/e-commerce/marketplace/mobile app].

Our business model: [e.g. $29/month subscription, freemium conversion to paid plan]
Main business metrics: [e.g. MRR or number of active subscribers]

Build a metric tree: Decompose the main metric into product components and product metrics into UX metrics that a designer can measure and improve.

Show the tree as a structure with indentations. For each UX metric, specify how to measure it (tool or method).

Claude will bring the tree back to your particular business model – not an abstract template.

### Prompt: Finding weaknesses in the current tree

Here is our current tree metric:
[Insert tree]

Find in it:
1. Metrics without a clear link to business results
2. Levels with gaps (something important is not measured)
3. Metrics that are easy to "gaming" (grow well without real product improvement)

Suggest specific corrections.

### Prompt: Translate the design task into the language of the metric

Often the designer knows what he wants to do, but he does not know how to justify the business.

I want to remake [the specific screen or float].

Help me build my argument through metrics:
1. What UX metrics should this improve?
2. How does this UX metric relate to product KPIs?
3. How do product KPIs relate to business metrics?
4. Suggest a wording to talk to the CEO or CPO.

Context of the product: [describe briefly]

### How to Use AI to Regularly Review a Tree

Once a quarter, ask Claude:

Our focus has changed: [for example, we used to be in the growth stage, now we move to monetization / launch B2B sales / enter a new market].

Here's the current tree metric:
[Insert tree]

What metrics should be added, removed or changed to a new focus? Explain the logic of each change.

It takes 10 minutes instead of a half-day workshop.