MM03 - Build My Frame Instrumentation
MM-03 · Full-Test Instrumentation Read
Build My Frame Funnel
Every step of the Build My Frame customizer now reports data. That lets us see four things at once: where people leave, which steps are hard to understand or execute, what shoppers are buying for, and how mobile and desktop behave differently. The customizer has two entry paths, the PDP flow and the homepage flow, tracked separately below.
The overview
Where people leave
Percentages are of the 10,230 who opened the builder over the full test. At Step 2 the flow splits three ways: measure now, buy now and measure later, or leave. Red marks people who leave. The purchase line at the bottom comes from the store's order log.
Open builder
10,230
100%
Step 2 decision · what the 10,230 openers do next
Measure nowcontinues the flow
5,843 · 57%
Buy Now, Measure Laterbranches to buy
1,103 · 11%
Left before choosingthe true leak
~3,284 · 32%
Measure width & heightStep 3, Continue
4,479
44%-23%
Space around the mirrorStep 4, Continue
3,877
38%-13%
Clips or channelsStep 5, Continue
3,225
32%-17%
Review & add to cartFinal step
2,848
28%-12%
Frame purchasesShopify, sitewide, both entries
2,542
see note below
Reading the whole flow
What each step tells us
Because every element reports data, we can separate who reached a step, who tried it, and who completed it. That is how we tell a hard step apart from a step people simply leave on.
Step 1
What size mirror are you framing?
Business signal
9,300pick a size
1,434skip the question
91%tell us their room
Four size tiles double as a survey of what people buy for. Double Vanity (3,495) and Single Vanity (2,824) dominate, XL (1,980) beats Powder Room (1,001). Most shoppers are framing mid-to-large vanity mirrors, and 59% pick the two priciest tiers. This is also the first price exposure ($99 to $260), and it barely deters anyone. See the room mix below.
Step 2
Are you ready to measure your mirror?
Biggest leak
5,843measure now (57%)
1,103buy now, measure later (11%)
~3,284leave (32%)
The largest single drop, and it lands the moment measuring becomes real. Measurement anxiety, not the builder itself, is the core problem. The Buy Now, Measure Later escape already catches 1,103 people who would otherwise likely leave, evidence the offer earns interest and deserves a more prominent place here.
Step 3
Measure your mirror's width and height
Hardest step
5,488engage the inputs (94%)
4,479continue (82%)
~1,009try but can't finish
This is where measuring actually breaks. Nearly everyone who commits tries to enter dimensions, but nearly 1 in 5 who start can't finish. The whole-inch plus fraction inputs and the "include the metal strip" tip are the most demanding entry in the flow. The demo video and FAQ live here, yet the FAQ is opened only 122 times across the whole test, so people push through or leave rather than look for help. Making the first measurement clearer is the highest-value fix.
Step 4
Measure the space around your mirror
Runs smooth
3,990engage (89%)
3,877continue (97% of engagers)
239escape to measure later
Once past the first measurement, the obstruction and clearance step is easy: 97% of people who engage it continue. The Top, Left, Right, Bottom dropdowns do not confuse anyone. This contradicts the earlier assumption that the middle of the flow was the weak point. The one wrinkle: measure-later escapes tick up here (159 at the first measurement to 239 here), a sign of fatigue rather than confusion.
Step 5
Does your mirror have any clips or channels?
Friction point
53%of buyers have clips
3,225continue (83%)
652drop here (17%)
Clips are not an edge case. 53% of completed frames have at least one clip or channel, averaging 1.8 sides, so this step does real work for half of buyers. It also costs 17% of the people who reach it. A clearer, faster clips question, with plain visuals of what a clip or channel looks like, is a strong test candidate.
Final step
Review your frame, then add to cart
End-stage drop
3,225reach the review
2,848add to cart
377drop at the end (12%)
377 people finish configuring their frame and then don't add to cart. The review screen is where the full price and every measurement appear together, so this is a moment of price reveal and last-second doubt about whether the numbers are right. One honest caveat: this drop was heavily front-loaded, 27% in the first week and far lower for the rest of the run, so read the 12% as the fair full-test rate. Reassurance at review, a confident price summary and a clear "your measurements look good" cue, remains a worthwhile end-stage test.
After the cart
Closing the loop with the order log
Purchases
2,542frame orders in the same 26 days
$743,600frame revenue
$293average frame order
A frame can only be bought through the builder, so the store's own order log closes the funnel: 2,542 frame orders were placed during the test (business accounts, phone orders and cancelled orders removed). That covers both builder entries and the measure-later branch combined, and the order log cannot split an order by entry point. This product-page flow holds about 8 in 10 of all builder add-to-carts, so on that split roughly 2,050 of these orders trace to it: about 1 in 5 of the 10,230 who opened the builder, and about 7 in 10 of the 2,848 who added to cart. Within the total, 122 orders still have measurements pending and 47 more completed their measurements after ordering.
Across the flow
Help-seeking is low, right where it's needed most
399click Need help (3.9%)
122open the FAQ (1.2%)
Very few people ask for help, yet ~1,009 abandon the first measurement. The help content exists but isn't catching the people who need it. Proactive, inline guidance at the first measurement, rather than a help link people have to go find, is worth testing.
Desktop vs mobile
The same flow, two very different experiences
Mobile is 78% of this flow's builder starts but converts far worse. Bars show each device's share of its own builder-starts, so the shapes are directly comparable.
Mobile · 7,973 starts
Desktop · 2,257 starts
Open builder
Commit to measuringStep 2, Measure now
Finish first measurementStep 3
Space aroundStep 4
Clips or channelsStep 5
Add to cart
Desktop completes 48%, mobile just 22%, a 2.2x gap. The split happens at the first measurement: desktop barely drops from Step 2 to Step 3 (69% to 67%), while mobile falls off a cliff (54% to 37%). Entering width and height with fraction dropdowns is manageable on a desktop and painful on a phone. Since mobile is 78% of the traffic, fixing the mobile measurement experience is the single biggest lever on this flow. Part of the gap also reflects a browse-on-mobile, buy-on-desktop pattern: site-wide over the last 30 days, desktop is just 19% of sessions but 57% of revenue, and its average order is about 50% higher than mobile's. High-value, measure-first purchases concentrate on desktop. So a second lever is to let mobile shoppers save or send their build and finish on desktop, rather than only trying to close them on the phone.
Beyond the funnel
What shoppers are buying for
Step 1's size tiles are a live read on demand, useful well beyond this test, for imagery, ad creative, and merchandising.
Room mix at Step 1
Share of the 9,300 shoppers who pick a size. Skip shown separately.
Double Vanity up to 6 ft, $225
3,495 · 38%
Single Vanity up to 4 ft, $185
2,824 · 30%
XL Vanity 6 ft+, $260
1,980 · 21%
Powder Room up to 2 ft, $99
1,001 · 11%
Skipped the question
1,434
- Feature vanities, especially double. Double and Single Vanity are 68% of selections. Lead imagery and ad creative with double and single vanity room scenes, not powder rooms.
- Buyers skew large and higher-value. 59% pick the two priciest tiers (Double and XL). The audience self-selects into the $185 to $260 band, a healthy AOV signal.
- Powder rooms are the smallest segment (11%). A minor use case, not the hero.
- Buy Now, Measure Later draws interest but rarely closes. 1,501 clicks across the test, and about 170 orders came through the measure-later path, roughly 1 in 9. High intent, low completion, and a clear opportunity to fix the measure-later purchase path.
From insight to action
Tests this points to
Each test is anchored to a specific number above, ranked by expected impact on frame conversion. Several build directly on ideas the MirrorMate team raised. These are initial directions, to be sharpened and cross-referenced with other customer data before they go into a roadmap.
Test 1
Rebuild the measurement inputs for mobile
Highest impact
Why: mobile completes just 22% vs desktop's 48%, and it collapses at the first measurement (54% to 37% vs desktop's 69% to 67%). Mobile is 78% of the traffic. Test: mobile-native inputs, large steppers instead of tiny fraction dropdowns, an inline visual guide, one field at a time. Read on: mobile Step 3 continue rate and mobile conversion.
Test 2
Calm the "ready to measure?" moment, with an escape to learn
High impact
Why: 32% leave at Step 2, the single biggest drop, the instant measuring becomes real. Test: reassure before the ask, a "most people finish in under 5 minutes" cue and a progress indicator, plus a clear "how it works" or "what's in the kit" link so early-journey users can get their questions answered instead of leaving (raised by the team). Read on: Step 2 measure-now rate and builder completion.
Test 3
Make the first measurement clearer: dimensions and FAQs inline
High impact
Why: ~1,009 people start entering width and height and can't finish, and the FAQ goes almost unused. The team asked for a clear product-details section that states frame dimensions and answers the top questions. Test: surface dimensions and the top FAQs inline at the measurement step, with a more prominent how-to-measure demo. Read on: Step 3 continue rate.
Test 4
Fix the Buy Now, Measure Later path
Clear opportunity
Why: 1,501 clicks but only about 170 completed orders across the test. The intent is there, the purchase isn't landing. Test: streamline the measure-later checkout and frame it as a real offer, stock reserved, cancel any time. Read on: click-to-order rate on the measure-later path.
Test 5
Lead with vanity imagery
Merchandising
Why: 68% of builders are framing single or double vanities and 59% pick the two priciest tiers. The team suggested leading with a framed mirror instead of a frame corner. Test: feature a framed double or single vanity scene as the first image, and swap the frame on a consistent bathroom photo as the shopper changes style. Read on: builder starts and frame conversion.
Test 6
Bridge mobile to desktop: save and resume your build
Opportunity
Why: desktop is 19% of sessions but 57% of revenue, with an average order about 50% higher than mobile's. High-value buyers finish on desktop, so a mobile shopper who leaves is not always lost, they may intend to complete elsewhere. Test: let shoppers save their build and pick it up on another device, a "save my build" or "email me this frame" option that returns them to the exact configuration. Read on: builds started on mobile and completed on desktop, and return-to-purchase rate.
The overview
Getting into the builder
The homepage customizer is the style-first version opened from the second fold. The top of the funnel, through the measurement commit, reports cleanly. Percentages are of the 4,284 who opened the builder over the full test.
Open buildersecond-fold Build My Frame
4,284
100%
Select a frameStep 1, style-first
3,037
71%-29%
Choose a sizeStep 2
2,602
61%-14%
Commit to measuringStep 3, Measure now
1,318
31%-49%
Instrumentation note
Why the homepage read stops at the measurement commit
- The top of the funnel reads cleanly, from opening the builder through the Step 3 measurement commit. That is the part reported above and below.
- The steps after the measurement commit have a tracking issue we caught in our own QA. One event counts every measurement dropdown touch rather than each person (3,573 interactions from the 1,318 who reached the step), and the Step 4 continue action is currently bound to the wrong button. So a reliable step-by-step read past this point is not available yet.
The dev team has the exact one-line fix. To keep the current roadmap on schedule we are not restarting this test, so this homepage read leans on the reliable top funnel and the device split. Both point to the same conclusion as the PDP flow: measuring on mobile is where it breaks.
Desktop vs mobile
The homepage is a mobile problem
The readable top of the funnel tracks closely by device, yet the overall conversion gap is even wider than on the PDP flow. Bars show each device's share of its own builder-opens.
Mobile · 2,502 opens
Desktop · 1,782 opens
Open builder
Select a frameStep 1
Choose a sizeStep 2
Commit to measuringStep 3, Measure now
| Metric | Mobile | Desktop | Gap |
|---|---|---|---|
| Share of homepage traffic | 83% | 17% | · |
| Open the builder | 3.8% | 13.6% | 3.5x |
| Conversion rate | 1.98% | 8.19% | 4.1x |
| Add-to-cart rate | 3.63% | 11.93% | 3.3x |
83% of homepage traffic is mobile, and it converts at 1.98%, less than a quarter of desktop's 8.19%. Desktop even opens the builder 3.5x more often per visitor, so of the 4,284 opens the split is 2,502 mobile to 1,782 desktop. The two devices track together through frame selection, then separate: mobile reaches the measurement commit at 25% against desktop's 38%, and the 4.1x conversion gap widens further in the measurement steps and checkout. Same root cause as the PDP flow: measuring on a phone is hard.
From insight to action
What Build My Frame needs before the redirect
Each is a test in its own right, mapped to a number above or to feedback the MirrorMate team raised. These are initial directions, to be sharpened and cross-referenced with other customer data before they go into a roadmap.
Test 1
Make Build My Frame mobile-first
Highest impact
Why: homepage mobile converts at 1.98% and is 83% of the traffic the banner would send. Test: the mobile measurement rebuild from the PDP tab, applied here first, since this is where the redirect would land. Read on: homepage mobile conversion.
Test 2
Add clear product details and up-front pricing
High impact
Why: the team gets frequent "how does this work" and "what does it cost" questions, and the page currently leads with the powder-room price, the cheapest tier and only 11% of picks, with no "starts at". Test: a clear details section with frame dimensions and top FAQs, plus a size-based pricing block that reads as "starts at". Read on: Step 1 to Step 3 progression and builder completion.
Test 3
Give cold visitors a way to learn and to trust
Readiness gate
Why: cold traffic needs to understand the product and see proof before committing, and the team specifically flagged the loss of the how-it-works content and reviews. Test: a visible "how it works" or "what's in the kit" escape and a customer-reviews section on scroll. Read on: builder completion for new visitors. Reviews impact will need its own read.
Test 4
Then test the "Shop Now" to Build My Frame redirect
Sequenced
Why: getting people into the measurement flow faster is a real opportunity, but only once the flow can hold cold, mobile traffic. Test: the redirect, run after Tests 1 to 3 land, measured as soon as the sample is significant. Read on: homepage builder starts and overall conversion.
Highlights from the full test
- 28% of the 10,230 who open the builder reach Add to Cart (2,848). Nearly three in four leave inside the flow, and the order log closes the loop: about 1 in 5 openers ends in a frame purchase.
- Mobile is where it breaks. Mobile completes 22% vs desktop's 48%, collapsing at the first measurement, and mobile is 78% of builder starts. The mobile measurement experience is the biggest lever we have.
- The biggest leak is before measuring even starts. At Step 2, 57% choose Measure now and 11% branch to Buy Now, Measure Later, but about 3,284 people (32%) leave without choosing either.
- The first measurement is the hardest step, not clips. ~1,009 people start entering width and height and can't finish. Step 4 (space around), by contrast, runs smooth at 97% completion.
- Step 1 is a merchandising signal: Double and Single Vanity are 68% of selections, and 59% pick the two priciest tiers. Feature vanities, not powder rooms.
- Buy Now, Measure Later draws clicks but rarely closes: 1,501 clicks, about 170 orders, roughly 1 in 9. Fixing that purchase path is a clear win.
- The review screen loses 377 people at the finish (12% across the full test, front-loaded into the first week), where full price and all measurements appear at once. Still a worthwhile late-funnel test surface.
- The homepage entry is a mobile story too: 83% of its traffic is mobile, converting at 1.98% vs desktop's 8.19%, and desktop opens the builder 3.5x more often per visitor. The reliable top funnel and device split are in; the steps past the measurement commit await a tracking fix.