Funnel Analysis in Practice: Breaking Down Conversions with HelloFresh

Funnel analysis is one of the most fundamental and practical methods in product analytics. It tracks user behavior through a sequence of ordered steps, quantifying conversion and drop-off at each stage. This article uses HelloFresh's subscription business model to walk through the full framework: defining funnels, calculating metrics, identifying bottlenecks, and developing optimization strategies.

Note: All numbers in this article are hypothetical, used solely to illustrate the analytical approach.

What Is Funnel Analysis?

Funnel analysis is a sequential model that tracks users from an initial touchpoint to a final goal (e.g., purchase, signup), measuring completion and drop-off rates at every step.

Core concepts:

• Conversion Rate: The percentage of users who complete the current step and move to the next
• Drop-off Rate: The percentage who leave at the current step without progressing
• Overall Conversion Rate: The end-to-end ratio from funnel top to bottom

How it differs from other methods:

Method	Question It Answers
Funnel Analysis	Where do users drop off the most?
Cohort Analysis	How do different user batches behave?
Retention Analysis	Do users keep coming back?

These three are often used together, but funnel analysis is your first tool for diagnosing conversion problems.

HelloFresh: Business Context

HelloFresh is the world's largest meal-kit subscription service. The business model is straightforward: users choose a meal plan online, receive ingredients and recipes weekly, and cook at home.

This model is ideal for demonstrating funnel analysis because:

1. Clear path: There's a well-defined linear flow from browsing to subscribing
2. Subscription mechanics: You track not just first purchase but also retention and reorders
3. Multiple friction points: Plan selection, signup, menu customization, and checkout each present potential drop-off

Key business metrics:

• CAC (Customer Acquisition Cost): Cost to acquire one new customer
• Trial-to-Paid Conversion: Rate at which trial users become paying subscribers
• Churn Rate: Monthly subscription cancellation rate

Defining Funnel Steps

A typical HelloFresh new-user funnel:

Step	Event Name	Description
Step 1	page_view_landing	Visit landing or campaign page
Step 2	plan_browse	Browse the plans page
Step 3	plan_select	Select a specific plan
Step 4	signup_complete	Complete registration or login
Step 5	menu_customize	Customize the weekly menu
Step 6	checkout_complete	Complete checkout and payment
Step 7	reorder	Reorder in week two or later

Granularity Trade-offs

More steps isn't always better:

• Too few (e.g., just "visit → checkout"): You can't see where the problem is
• Too many (e.g., splitting every form field): Too much noise, hard to focus
• Recommended: 3–7 steps, each corresponding to a meaningful user decision point

Calculating Conversion and Drop-off Rates

Assume we tracked one month of data:

Step	Users	Step Conversion	Cumulative Conversion
Landing Page	100,000	—	100%
Plan Browse	60,000	60.0%	60.0%
Plan Select	25,000	41.7%	25.0%
Signup	15,000	60.0%	15.0%
Menu Customize	10,000	66.7%	10.0%
Checkout	7,000	70.0%	7.0%
Reorder	3,500	50.0%	3.5%

Formulas:

Step Conversion = Users at current step / Users at previous step
Cumulative Conversion = Users at current step / Users at step 1
Drop-off Rate = 1 - Step Conversion

Quick takeaway: Plan Browse → Plan Select (41.7%) and Checkout → Reorder (50.0%) are the two biggest bottlenecks.

Finding Bottlenecks: Common Analytical Techniques

Looking at the overall funnel isn't enough — you need to slice by dimensions to diagnose root causes.

1. Step-by-Step Drop-off Analysis

Identify the step with the largest drop-off and prioritize it. In our example, "Plan Browse → Plan Select" loses 58.3% of users — that's where to investigate first.

2. Segmentation by Channel

Channel	Landing → Checkout Conversion
Google Ads	5.2%
Organic Search	8.1%
Referral	12.3%
Social Media	3.8%

Referral conversion is significantly higher than paid ads, suggesting the referral program may deliver better ROI.

3. Segmentation by Device

Device	Menu Customize → Checkout Conversion
Desktop	78.5%
Mobile	58.2%

Mobile checkout conversion is notably lower — likely a UX issue in the mobile checkout flow.

4. Time-to-Convert Analysis

How long do users spend at each step before moving on?

• Plan Browse → Plan Select: median 4.2 minutes (long deliberation — plan comparison may be too complex)
• Menu Customize → Checkout: median 1.1 minutes (smooth, no obvious friction)

Extended dwell time usually signals user confusion or hesitation.

Optimization Strategies

For each bottleneck, apply targeted optimizations:

Browse → Select (41.7%, biggest bottleneck)

• Simplify plan comparison: Reduce the number of plans, highlight differences, add recommendation labels ("Most Popular")
• A/B test the pricing page: Test different layouts, default selections, and price anchoring
• Add social proof: Display subscriber counts, ratings, and user reviews

Select → Signup (60.0%)

• Reduce signup friction: Support Google/Apple social login
• Defer registration: Let users finish menu selection before requiring signup (create sunk cost first)
• Guest checkout: Allow purchase without account creation

Menu Customize → Checkout (70.0%)

• Default menus: Offer system-recommended default selections to reduce choice burden
• Save preferences: Remember dietary preferences (vegetarian, gluten-free) and auto-filter

Checkout → Reorder (50.0%, retention key)

• Onboarding emails: Send cooking tips and recipe videos during the first week
• Difficulty matching: Recommend recipes matching the user's cooking experience
• Flexible pause: Allow skipping a week instead of outright cancellation
• Exit retention: Offer discounts or plan adjustments during the cancellation flow

Prioritization: Impact × Effort

Optimization	Expected Impact	Implementation Effort	Priority
Simplify plan comparison	High	Low	P0
Social login	Medium	Low	P1
Default menus	Medium	Medium	P1
Onboarding emails	High	Medium	P0
Flexible pause	High	High	P2

Tools and Implementation

Analytics Platforms

• Mixpanel / Amplitude: Built-in funnel analysis with drag-and-drop step configuration
• Google Analytics 4: Funnel exploration report (Exploration → Funnel)
• Custom solution: Event tables + SQL queries

SQL Example

Given an events table events(user_id, event_name, timestamp), use window functions to build the funnel:

WITH funnel AS (
  SELECT
    user_id,
    MAX(CASE WHEN event_name = 'page_view_landing' THEN 1 ELSE 0 END) AS step1,
    MAX(CASE WHEN event_name = 'plan_browse' THEN 1 ELSE 0 END) AS step2,
    MAX(CASE WHEN event_name = 'plan_select' THEN 1 ELSE 0 END) AS step3,
    MAX(CASE WHEN event_name = 'signup_complete' THEN 1 ELSE 0 END) AS step4,
    MAX(CASE WHEN event_name = 'checkout_complete' THEN 1 ELSE 0 END) AS step5
  FROM events
  WHERE timestamp BETWEEN '2026-02-01' AND '2026-02-28'
  GROUP BY user_id
)
SELECT
  COUNT(*) AS total_users,
  SUM(step1) AS landing,
  SUM(CASE WHEN step1 = 1 AND step2 = 1 THEN 1 ELSE 0 END) AS browse,
  SUM(CASE WHEN step1 = 1 AND step2 = 1 AND step3 = 1 THEN 1 ELSE 0 END) AS selected,
  SUM(CASE WHEN step1 = 1 AND step2 = 1 AND step3 = 1 AND step4 = 1 THEN 1 ELSE 0 END) AS signup,
  SUM(CASE WHEN step1 = 1 AND step2 = 1 AND step3 = 1 AND step4 = 1 AND step5 = 1 THEN 1 ELSE 0 END) AS checkout
FROM funnel;

This SQL ensures users must complete steps in order to be counted at each stage, preventing step-skipping from inflating numbers.

Presenting to Stakeholders

• Use a horizontal bar chart for the funnel, annotating each step with conversion rates and absolute numbers
• Add a trend line chart tracking weekly funnel conversion changes
• Include a segmentation comparison table so product managers can instantly spot which channel or device needs attention

Common Pitfalls

1. Vanity Metrics Only

Optimizing only the top of the funnel (e.g., landing page traffic) while ignoring downstream conversion won't drive revenue growth.

2. Unclear Attribution Windows

A user browses today but checks out next week — does that count as the same funnel? You need a well-defined attribution window (e.g., all steps completed within 7 days).

3. Survivorship Bias

Analyzing only users who entered the funnel ignores those who never entered at all. Perhaps the biggest problem isn't inside the funnel — it's that nobody knows your product exists.

4. Over-segmentation

Slicing data across too many dimensions (channel × device × region × new vs. returning) shrinks sample sizes and makes conclusions unreliable. Verify each segment has sufficient sample size before drawing conclusions.

5. Numbers Without Context

Funnel analysis tells you where the problem is, not why. Pair it with qualitative methods — user interviews, session recordings, surveys — to find the real root cause.

Conclusion

Funnel analysis is a starting point for product analytics, not the destination. It helps you quickly pinpoint conversion bottlenecks, but real optimization requires combining segmentation, A/B testing, and qualitative research.

Using HelloFresh as our example, we can see that even a seemingly simple "browse → buy" flow contains multiple optimization opportunities when decomposed. The key is: define clear steps, track the right events, use data to find bottlenecks, then prioritize with Impact × Effort to decide what to tackle first.

Remember: funnels evolve as your product evolves. Periodically revisit your funnel definitions to ensure they reflect the current user journey, not assumptions from six months ago.