FICTION · An Unfair Advantage Intervention
Fiction. Real history. Sourced below.
Everyone agrees with the Zappos lesson. Find the thing stopping the purchase and remove it. What the case actually demonstrates is how few organisations can hold that position once the cost arrives at scale, because the industry that copied it has spent the last few years adding return fees, and Zappos itself reversed the culture experiment it was most famous for.
This is a review of a service model, written the way I would write one for a board considering something similar. I am Samantha Reynolds and I built Cartwell’s incentive system, which is the lens I bring: I do not read the Zappos call centre as a service story. I read it as a measurement decision, and the measurement decision is the part that actually transferred to anybody else.
Full fiction disclosure and sourcing note
Samantha Reynolds is a fictional character from The Unfair Advantage by Todd Hagopian, and Cartwell Manufacturing is the book’s fictional company. She analyzes the documented history of Zappos from the present day. She has never met, advised, or corresponded with anyone named here, and no dialogue, thought, or motive is invented for any real person. Facts come from the public record and are accurate as of 2026; the analysis is invented. Contested points are flagged: published accounts of Zappos’s current return window disagree, with some reporting the long-standing 365-day policy intact and others reporting a 60-day refund window with store credit thereafter, and the customer-lifetime-value figures often quoted for Zappos are illustrative rather than disclosed. Tony Hsieh’s death is stated as a matter of record and is not treated as part of the analysis.
- The Founding Insight Was a Statistic, Not a Vision
- The Conversion Death Point Was Return Risk
- The Call Centre Was an Incentive System First
- The Consensus Is Right, Which Is Why the Retreat Matters
- The Culture Experiment Was Reversed
- The Contrarian Reappraisal
- About the Stagnation Assassin
The Founding Insight Was a Statistic, Not a Vision
Zappos was founded in 1999 by Nick Swinmurn, initially as ShoeSite.com, after he failed to find a pair of shoes at a mall and then failed to find them online. Tony Hsieh, who became the company’s public face, was an investor first, through his fund Venture Frogs, joining as co-chief executive around 2000. Swinmurn left in 2006.
The artifact worth examining is not a policy. It is the number Swinmurn used in his pitch: footwear was roughly a $40 billion market in the United States, and about five percent of it was already being sold through mail-order catalogues. Sit with what that establishes. One shoe purchase in twenty was already being made by people who could not try the product on. The trust barrier was not hypothetical and it was not absolute. It was measurable, and it was already being crossed by a small population under worse conditions than an internet retailer could offer.
That reframes the founding from vision to observation. Swinmurn did not intuit that people would buy shoes sight-unseen. He found evidence that some already did, and reasoned that the constraint on the rest was the cost of being wrong rather than the fact of not trying them on. Every subsequent policy follows from that read. A company that believed the barrier was psychological would have invested in reassurance and imagery. A company that believed it was economic invests in removing the cost of a mistake, which is what happened.
The Conversion Death Point Was Return Risk
Magnificent Obsessions has a customer half concerned with where purchases actually fail rather than where companies assume they do, and its sharpest instrument is the Conversion Death Point: the specific moment at which an intended purchase stops. Identify that point precisely and you often find that the highest-leverage move is not improving the things you already do adequately.
For online shoes in 1999 the death point was not price, selection, or site design. It was the instant a customer imagined the shoes not fitting and calculated what that would cost them: return postage, a wait for the refund, possibly a restocking fee, and the argument required to obtain it. Zappos removed all four. Free shipping outbound, free shipping back, a return window measured in months rather than weeks, no restocking charge, no dispute.
The result was an inversion rather than an improvement. With returns genuinely free, ordering three sizes online and keeping one became lower risk than visiting a store, because the customer could try the shoes on their own floor with their own clothes and return the rest at no cost. The company did not compete on a dimension its rivals were contesting. It removed the dimension.
Worth noting for anyone applying the framework: Magnificent Obsessions is a HOT System instrument, built for crisis conditions, and Zappos was a young company in none of the three doctrinal conditions. The tool travels here. The doctrine does not apply.
The Call Centre Was an Incentive System First
This is the part I want to claim for my own discipline, because it is routinely misfiled. The famous Zappos service behaviours, representatives with no script, calls running for hours, agents directing a customer to a competitor when an item was out of stock, are not a culture outcome. They are what happens when you change what is measured.
A conventional contact centre measures handle time, calls per hour, and adherence. Those metrics reliably produce short calls, because that is what they reward, and every representative learns within a fortnight which behaviours are counted. Zappos measured whether the customer was satisfied and removed the volume metrics. Everything downstream is a rational response by ordinary people to a different scoreboard, which is why the behaviours were consistent across thousands of employees rather than concentrated in a few unusually motivated ones.
That distinction matters for anyone copying this. You cannot instruct people into it, and a values statement will not produce it. If handle time is still on the dashboard, the ten-hour call does not happen regardless of what the poster in the break room says. The transferable act is deleting a metric, which is cheap to do and organisationally difficult, because the metric exists to control a cost that somebody is accountable for.
The Consensus Is Right, Which Is Why the Retreat Matters
Here is the shape of this case that I think is usually missed. There is no serious disagreement about the insight. Remove the binding barrier, measure lifetime value rather than transaction margin, treat service as a channel: every retail executive endorses all three, and has for fifteen years. The interesting question is therefore not whether the lesson is correct. It is why so few sustained it.
The answer is that the cost is real and it grows with volume. Returns in footwear and apparel are expensive, and as e-commerce scaled through the last decade the aggregate cost stopped being a rounding error and became a line item large enough to change behaviour. Much of the industry has since introduced return fees, shortened windows, and moved against serial returners, which is precisely the abuse the confident version of this lesson said was safe to absorb.
Zappos’s own current policy is now genuinely unclear from the outside, and I flag that rather than resolve it. Some 2026 accounts report the 365-day window intact; at least one, checking the company’s published policy in August, reports a 60-day refund window with store credit available thereafter. If the second is right, then the single most imitated policy in modern retail has been quietly shortened by the company that made it famous, which would be the most informative fact in the entire case.
None of this makes the original insight wrong. It makes it conditional in a way the slogan version deletes. The model worked because shoes are fit-dependent, repeat-purchase, and carry margins that could fund the returns. Strip any of those and the same generosity is just a subsidy with no recovery mechanism.
The Culture Experiment Was Reversed
The other thing the inspirational telling freezes out is what happened to the organisation. From 2013 Zappos moved to holacracy, a self-management structure without managers or job titles, and in March 2015 employees were given a memo asking them to commit or take severance. Around 210 of roughly 1,500 employees, about 14 percent, took the first offer, with a second round taking the total to roughly 18 percent. Around 300 more left through buyouts in 2020.
Then it was undone. Zappos quietly moved away from holacracy, restoring managers while keeping some structural artifacts, and John Bunch, who co-led the rollout, attributed the change to big challenges in the company’s business metrics and a need to point employees back at the customer. A recurring criticism of the model is that it directs attention inward, which is an unfortunate outcome for a company whose entire advantage was outward attention.
Amazon, which acquired Zappos in 2009 for around $1.2 billion in stock and initially left it alone, has since absorbed functions and standardised operations, down to replacing the distinctive branded packaging. Tony Hsieh stepped down in August 2020 after 21 years and died that November at the age of 46. I state that as a matter of record and will not make it part of an argument about retail strategy.
The Contrarian Reappraisal
I have argued that the barrier-removal insight is correct and rarely sustainable. There is a stronger objection than the one I have been answering, which is that the returns policy was never the innovation at all.
On this reading, liberal returns were not unique to Zappos even in 2005, catalogue retailers had offered generous terms for decades, and Amazon was already building a reputation for frictionless refunds. What Zappos actually sold was assortment, a genuinely deep inventory of sizes and widths that physical stores could not hold, wrapped in service that made ordering multiples feel normal. The policy was the permission slip. The stock was the product. If that is right, the industry’s retreat from free returns is not a failure to sustain a hard discipline; it is the correction of a temporary condition in which e-commerce was buying customer acquisition with return subsidies, and Zappos simply did it earlier and more gracefully than anyone else.
I find that argument about half persuasive, and the half I accept is the assortment point. Depth of inventory was a real and underrated part of the offer, and the reason ordering three sizes worked is that three sizes existed to order. What the objection cannot explain is the incentive design. Deep stock is a purchasing decision any capitalised competitor could match, and several tried. Removing handle-time metrics from a contact centre is nearly free and almost nobody did it, which suggests the constraint was never capital. It was the willingness to stop measuring a cost.
So the version I would hand a board is narrower than the legend and less dismissive than the objection. Find where the purchase actually dies, which requires evidence rather than assumption. Check whether your category can fund removing it, which is arithmetic and not conviction. And if you want the behaviour, change the scoreboard rather than the poster, because that is the part of this case that cost the least and travelled the furthest.
About the Stagnation Assassin
Todd Hagopian is a Fortune 500 product and transformation executive who has driven $3B+ in shareholder value across Berkshire Hathaway, Illinois Tool Works, Whirlpool, and JBT Marel, where he is VP of Product Strategy and Marketing. Known as The Stagnation Assassin, he writes on customer strategy, incentive design, and the difference between a lesson executives endorse and one they can actually sustain. He turns proprietary frameworks including Magnificent Obsessions, Revenue Responsibility Engineering, and the HOT System into field-tested tools for executives, and is the author of the Stagnation Assassin series, a Forbes contributor, and an award-winning keynote speaker. Connect with Todd on LinkedIn or visit his speaking page.
If handle time is still on the dashboard, the ten-hour call does not happen, whatever the values statement says. Change the scoreboard or change nothing.
Todd speaks to executive teams and conferences on customer strategy, incentive design, and finding where purchases actually die. Book a keynote.