On’s second pair: how I’d separate a replacement from a running rotation
By Trayan Hristov · Founder, Thrivelia · 19 November 2026
A runner orders the same shoe in the same size eight months later. Another buys a different pair three weeks after their first order. Both count as repeat buyers. One might be replacing a favourite. The other might be building a rotation. Either could also be buying a gift. A date and a SKU get us part of the way. They don’t tell us what the next shoe is for.
On’s product guides already distinguish shoes by intended use, including daily training and longer runs. Its 2025 filing reported 30% net sales growth and more wholesale than direct sales. That channel mix matters: an apparently first-time online customer might have owned the shoe for years. I’d label the data “first observed direct order” and give the customer an easy way to correct the picture. On’s Cloudmonster guide · On’s annual filing.
Start with the runner’s question
For one market and one road-running family, I’d test an optional choice in the owned journey: “Replacing a pair you like, looking for a different run, or buying for someone else?” A replacement buyer could see their earlier model, what changed in the current version, and which sizes are available. An intended rotation buyer could compare what their current shoe is meant for with one suitable option for a different run. A gift buyer gets help choosing for someone else. Skipping the question keeps the ordinary experience.
The asset I’d ask footwear merchandising to make is a side-by-side explanation of two use cases, in plain English. A runner deciding whether the second pair earns space in the hallway needs a reason for each shoe. A catalogue sorted by release date answers a different question. I wouldn’t infer mileage, injury or a replacement date from a previous order.
Count the first and the third observable orders properly
Compare eligible people assigned to the new choice with the current journey. Measure kept second direct footwear purchases over a pre-agreed window long enough for replacements. Use a shorter diagnostic window for expressed rotation intent. At order two, record whether it was a replacement, addition, gift or unknown. At order three, check whether rotation buyers later replace one of those pairs. Keep survey nonresponders in the assigned test groups.
Include returns, contribution per eligible buyer and stock availability. Retail purchases invisible to the test can make direct repurchase appear low. A change in the DTC result could also be channel switching. Ask retail to share an aggregate view where available before calling these extra brand sales.
Which Cloudmonster belongs next to the pair a runner already owns?
I’d ask one question before showing the alternatives: “Which run can’t your current pair do the way you want?” On positions the Cloudmonster 3 for everyday training and longer runs. The 3 Hyper also covers longer runs and adds tempo sessions and intervals. “I do long runs” still leaves two plausible choices. A person who just wants the familiar shoe in a wider fit has given us a much clearer answer. On’s Cloudmonster collection · On’s UK men’s range and widths.
A good reviewer can help a merchandiser explain that trade-off. Carlos Sánchez’s RunRepeat test of the 3 Hyper found it responsive and durable, and flagged the firm feel and weight as things some runners might dislike. I’d credit that test on a comparison card and let the runner decide whether that feel suits the session they have in mind. Carlos Sánchez’s Cloudmonster 3 Hyper test.
The runner’s answer | The useful product path | What could otherwise go wrong? |
“My everyday pair wore out.” | Show the current version, changes and stock in the buyer’s width. | An unavailable size looks like lost loyalty. |
“I want a pair for faster sessions.” | Compare the pair they own with the Hyper for that run. | A second long-run description adds little. |
“I need more room.” | Check the wide option in the right model and market. | A performance upgrade leaves the fit issue open. |
“It’s for someone else.” | Let them shop for the recipient without changing their own fit preferences. | A gift distorts the next recommendation. |
What order two can tell us about order three
Here’s the arithmetic I’d give the analyst. Imagine 200 buyers have kept their second direct order and all have another 180 days of observation. Of 80 who described that purchase as a replacement, 16 buy a third time. Of 40 who described it as a rotation, 16 buy again. Of 80 with unknown intent, eight return. That’s 20%, 40% and 10% in those groups, and 20% overall. Those are illustrative numbers.
It would be easy to conclude that getting more runners into a rotation doubles repeat. Runners who already need several pairs are more likely to choose that answer in the first place. I’d randomise the comparison card within the group looking for a second use, keep the ordinary shopping experience as control and count everyone assigned. If direct sales rise, retail sell-through and kept contribution help us judge how much demand we actually added. How stock availability can hide repeat demand.
The workshop I’d run with On
I’d bring a footwear merchandiser, retail analyst and CRM lead. Give them three anonymised baskets each and ask them to label the intended job independently. Where they disagree, ask whether the customer could answer in one tap. If runners already know their shoes and the real friction is finding their size, I’d put the experiment into stock availability instead.
Add the brand’s perspective
This is Thrivelia’s independent reading of public material. If you work at On, send a correction, a public source or a test you can share to office@thrivelia.com. Include this article’s title and say whether we may credit you. Use the brand response worksheet or read the method behind this test.

.png)



Comments