flowpeekresources / guides / Conversion
Conversion funnel: how to find the problem with mobile purchases
You'll go from weaker results on phones to a concrete check of the purchase process.
Short answer
The key answer
The funnel shows the scale of recorded purchase steps. Compare devices for the same period, pick a meaningful difference and verify it in user journeys and a hands-on test of the store. A drop between steps is a pointer for a check, not a ready-made diagnosis of a bug.
- Choosing the step that needs attention
- Comparing phones and computers together with visit counts
- A fix description based on a specific test
1. Check which actions make up the funnel.
The steps come from available e-commerce events: product view, product click, cart, checkout start, shipping and payment info, and purchase. Not every platform delivers every step in the same way. First find out which of these actions are actually collected in your setup.
The funnel lists people with a recorded step in the selected range. Don't read the table as a recording of one continuous visit. A customer can come back, and tracking may not cover some actions. To confirm the order of steps in a specific case, you need the user journey. The checkout form's URL alone doesn't prove that all its steps were completed.
2. Compare devices on a common basis.
Choose the same period and store, then read the breakdown by device. Mobile means phones, desktop means computers. Next to the percentages, note the number of people at each step. A difference of a few percentage points based on a handful of cases calls for different caution than a similar difference across many visits.
In the table, check product, cart, checkout start and purchase. The device filter lets you narrow the funnel; browser and source views are also available. Don't compare a promotion week on phones with an ordinary week on computers. Align the range first, then ask about causes.
3. Choose a place to check, not a full rebuild.
Many product viewers and few carts point you to the offer, variant availability, price or button visibility. A gap between cart and checkout start calls for checking shipping costs and how people move on. At the final steps, it's worth looking at forms and payment methods.
This is a list of hypotheses, not an automatic verdict. Before you order a fix, check that the step is recorded correctly. An event suddenly disappearing after a release can be a tracking change, even while customers keep buying. Pick one difference that matters for the store and note how you'll check it.
4. Narrow the test to the real environment.
The devices and browsers report helps set priorities. Combine the scale of traffic with sales, order value and conversion for that group. Then check the store on the selected phone and browser. Don't try to test every possible configuration with the same intensity — start with the environment that matters to your customers.
Note the product, variant, shipping and payment method used in the test. Look at the on-screen keyboard, error messages and elements covering the button. A difference in a metric between reports can stem from a different way of counting purchases or a different period; compare readings from the same view before you call a change an improvement.
5. Combine journeys, testing and the post-change result.
Review the available journeys for the selected step. Look for a recurring place where visits end or return to an earlier page. These examples help you prepare a test scenario, but on their own they don't describe the customer's motivation. Add a check of the store and, if you can, conversations with users.
The task description for your team should include the environment, the step, the observation and the expected behavior. After the release, go back to the same report and filters. Take traffic, promotions and the number of cases into account. If you changed the campaign and the checkout form at the same time, don't attribute the whole difference to one fix. Note what the result still doesn't settle.
Common mistakes
Every difference means a phone bug
Devices can attract different traffic and people at different decision stages. Check the context and how tracking works.
A percentage alone is enough
Also note the number of people at each step. A small sample and an incomplete period can produce a large apparent change.
For your next review
Checklist
- I've confirmed which steps are collected.
- I'm comparing the same period and store.
- I've noted the counts next to the percentages.
- I've chosen a device and browser for the test.
- I have a specific task description and a plan to measure again.
Now your store
Prepare a checkout review task
- 1
Pick one device group and one step.
- 2
Open a few available journeys and complete a test purchase.
- 3
Describe the problem so that someone else can reproduce it.
From report to decision
Practice example — illustrative values
A made-up set for learning. We don't copy results from the screenshots and don't assume the cause of the difference.
- Mobile
- 200 people with a cart, 80 who started checkout
- Desktop
- 100 people with a cart, 60 who started checkout
- Test
- Shipping cost and moving on from the cart on a phone
The ratios are 40% and 60% respectively. That's a reason to check a specific step, not proof that 40 more customers would have bought after a button change.
Questions and answers
Does every drop in the funnel mean lost sales?
No. Some people are still comparing offers, and some actions may be outside the tracking scope. The report helps you decide where to check.
Does the report confirm the order of steps for each person?
A step summary doesn't replace analyzing a specific journey. Check the order of available actions at the visit level.
What if I have few orders?
Extend the comparable period and pay attention to the scale of earlier steps. A hands-on store test can still reveal a problem you can't reliably judge from a percentage alone.
From knowledge to report
Open the right module.
Conversion funnel
The conversion funnel shows each step of the purchase: product, cart, checkout form and purchase. See how many visits reach each one and choose where to start making improvements.
Devices and browsers
See which devices and browsers customers use to visit your store and how often they buy. Pick the setup where you should test checkout first.
User journeys
Open a visit's history and follow the path from the first page to the cart or order. See where the visitor came from, which products they viewed and what they did, step by step.
Read the next guide.
All guidesAnalytics / 7 min
How to read the Flowpeek dashboard and pick your first topic to analyze
Learn the difference between a visitor, a page view and an order. Read each result against the period, your own goals and the detailed reports.
Sales / 10 min
How to analyze product sales in an online store
Connect views, carts, purchases and revenue. Learn the product report definitions, four analysis scenarios and a list of questions to ask before changing your offer.
Setup / 7 min
Setting up Flowpeek: how to prepare your store and sign off the setup
Gather your store details, check the first visit and order, agree on the data scope. A simple sign-off checklist — no self-installation needed.
Free · 30 minutes · online
GA4, your store and your ads show different numbers? We'll check it with you.
We compare one month of sales across your tools and point out three fixes to your measurement. No strings attached. What we check
30 days on your own data
Take a closer look at the page where the difference appears.
Once you've chosen where to check, use journeys and events. Prepare one change, then come back to the same ranges in the funnel.
- No fixed-term contract
- We set everything up for you in 48 h
- App servers in Germany (EU)
- Cancel anytime
30 days for PLN 0 · first payment after the trial · cancel anytime
See pricing after the trial