Kill a sales channel when its fully loaded contribution margin trails your next best channel for two consecutive quarters and no untested lever remains. For most Shopify brands under $2M, that channel is physical retail or wholesale, not paid media.
The channel a founder cannot bring themselves to close is almost always the channel the business started with.
Most merchants can name their best sales channel in under a second. Very few can name their worst one, and that gap is where margin quietly goes to die. The pattern shows up most often between $500K and $2M in annual revenue, where a business has usually accumulated a second or third place customers can buy, and nobody has stopped to ask what those places actually earn once every attached cost is counted.
The channel in question is rarely paid media, because paid media reports its own cost every day. It is the market stall, the wholesale account, the seasonal pop-up, the small retail lease. These channels report revenue clearly and cost vaguely, and the human brain treats a visible number as more real than an invisible one.
Whether you are doing $50K months or $500K months, the decision framework is the same. What changes is the size of the hole. This piece covers how to measure a channel honestly, what the evidence says about closing one, and how to exit without handing your customers to somebody else on the way out.
Merchants keep unprofitable channels open because channel revenue is visible in the dashboard and fully loaded channel cost is not. A booth that generates $4,800 over a weekend feels like a win, and the receipt for that win is immediate. The van hire, the two staff at weekend rates, the booth fee, the stock that came back unsellable, and the founder’s Saturday do not appear anywhere near that $4,800 in any report the merchant reads.
There is a second reason, and it is emotional rather than analytical. The legacy channel is often the one the business started with. Founders who built their first customer base at markets or through a single wholesale buyer carry a loyalty to that channel that survives long past its usefulness. Closing it feels like disowning the origin story.
The financial consequence is well documented across the Shopify ecosystem. A brand can post healthy top line numbers and still run negative contribution margin because different expense types hit the business differently, and almost nobody connects the two views in real time. That disconnect is the subject of a Fastlane conversation with Adam Callinan, who scaled Bottlekeeper past $60M with four people, on why a green ROAS can coexist with a shrinking bank balance. His framing applies directly here: every channel metric is a vanity metric until it is grounded in the P&L.
For operators under $500K, the stakes are usually founder time rather than cash. For operators above $2M, a legacy channel that consumes a full time person and returns 6% of revenue is a straightforward allocation error. Both versions are the same mistake measured in different currencies.
Calculate channel cost by subtracting every expense that would disappear if the channel closed, across a full twelve months, from that channel’s revenue over the same period. The twelve month window matters because seasonal channels flatter themselves over a quarter. A Christmas market looks extraordinary in December and irrelevant in February, and the annual view is the only one that tells the truth.
Start with the obvious costs: rent or pitch fees, staffing hours at the rate you actually pay, travel and vehicle costs, insurance, permits, and any hardware or fixtures amortised across their useful life. Then add the ones merchants routinely skip. Inventory shrinkage and damage at in person events runs materially higher than warehouse pick and pack. Payment processing on in person transactions carries its own rate. And the founder hours spent on that channel have an opportunity cost equal to whatever the next best use of those hours returns.
Shopify makes the revenue half of this straightforward if the channel is running through the platform. Shopify POS Lite is included with every plan and POS Pro carries a per location monthly fee, and either version tags in person orders to a location so channel level revenue can be pulled cleanly rather than reconstructed from memory. For brands above $1M who want the cost side connected to the revenue side automatically, a profit clarity tool such as Pentane joins accounting data to revenue and ad platform data so contribution margin by channel is a live number rather than a quarterly archaeology project.
The output of this exercise is one figure per channel: contribution margin per dollar of revenue. Rank your channels by that figure. Illustrative benchmark for brands in the $500K to $2M range: a healthy owned online channel typically clears a contribution margin several times that of a staffed physical channel operating at a comparable revenue level, because the physical channel carries labour that does not scale with volume.
BOTB spent close to two decades acquiring customers from staffed stands inside UK airports and shopping centres, then closed every one of them because rent and staffing costs at those sites were rising faster than the sites were returning. The company had started in 1999 by craning supercars into airport terminals, beginning at Heathrow, and grew that footprint across most major UK airports before reversing the decision entirely.
The reasoning the company published at the time is the useful part for operators. Rent and staff expenditure in retail locations climbed year on year, producing falling efficiency. A shrinking number of physically served customers was disproportionately constraining pricing strategy across the whole business. And digital channels had already been proven, through repeated trials, to execute the same acquisition job more effectively. The exit was not a reaction to a bad quarter. It was the conclusion of a multi year comparison between two channels that had been running side by side.
What followed is the part that gets left out of most channel exit conversations. Revenue for the six months to 31 October 2018, the period covering the transition, came in at £7.12 million against £5.54 million the year before, with adjusted profit before tax up 14.7%. The business did not shrink into its remaining channel. It grew, because closing the weaker channel let it tailor competitions, pricing, and product exclusively to the customer base that was actually compounding. Today the same business runs BOTB’s car competitions entirely online, and the group around it reported an oversubscribed AIM listing raising £40 million gross in its most recent audited full year results.
The transferable lesson has nothing to do with prize competitions. It is that a channel with two decades of history, a recognisable footprint, and real revenue attached to it was still the correct thing to close, and the business that closed it accelerated.
Run four checks on any channel you are unsure about: fully loaded contribution margin, customer overlap, founder time consumed, and whether an untested lever still exists. A channel that fails one signal is a fix candidate. A channel that fails three or four has already made the decision for you.
The customer overlap signal deserves the most attention because it is the one merchants misread. A channel that converts people who were already on your email list is not acquiring customers. It is relocating a purchase that would have happened anyway, at a higher cost to serve. Test this by matching the email addresses captured through the channel against your existing list. If the overlap runs high, the channel is a fulfilment method wearing an acquisition costume.
The fixability signal is the honest brake on premature closure. If you have never tested a different location, a different day, a different price, or a different product mix in that channel, you do not yet have enough evidence to close it. The practical guide to running that test cheaply sits in the Fastlane walkthrough on how to plan, cost, and evaluate a temporary retail experiment, which covers the postmortem metrics most brands skip. Run one clean test. If the numbers do not move, you have your answer.
A channel exit only pays off if the capital and attention it frees move somewhere with better unit economics, which for most Shopify brands means retention infrastructure rather than additional paid media. Closing a channel and pouring the savings into Meta simply relocates the same margin pressure to a more competitive auction.
The strongest post exit move is usually the one that raises revenue per existing customer. A three email post purchase sequence in Klaviyo or Omnisend, a review request around day 14, and a win back at day 45 will recover a meaningful share of one time buyers within 60 days, and none of it carries the marginal cost that a staffed channel does. Brands above $500K with a natural replenishment cycle can go further and layer a subscription on the highest repurchase SKU through Recharge, which converts a one off acquisition cost into a recurring revenue stream. The complete breakdown of subscription models, churn benchmarks, and app selection by stage is worth reading before committing to that path, because subscription is a system rather than a feature and it fails when treated as the latter.
The prize competition business referenced earlier followed the same logic. Having closed its physical footprint, it did not simply buy more traffic. It broadened what an online only model made possible, extending from cars into its house competitions and other prize categories, then in July 2025 launched a subscription pass that gives one recurring payment access to entries across every category. The sequence is the instructive part: close the channel, use the freed attention to deepen the surviving channel, then add a recurring layer on top of it.
For brands under $500K, the replacement is almost always email and a tighter product range rather than a new tool. For brands above $2M, the freed operator headcount is usually worth more than the freed cash. The stage by stage view of which retention lever to pull first maps this out by revenue band, and the ordering matters more than the tooling.
Wind a channel down across 60 to 90 days by capturing contact details first, announcing the closure directly to those customers second, and running a redirect offer before the channel goes dark. Closing abruptly hands every customer that channel acquired to whoever occupies the space next.
The capture step comes first because it is the only one that becomes impossible later. Every in person transaction should be attached to an email address or phone number before the final trading day, which Shopify POS handles natively by creating customer profiles at checkout. Illustrative benchmark: post purchase opt in rates at the point of sale run far ahead of generic website popups, because the customer has already bought and the exchange feels fair rather than extractive.
The announcement step is where most brands underperform by being vague. A direct message that names the closing date, explains plainly that the business is consolidating so it can serve customers better online, and includes a specific reason to place the next order through the website converts noticeably better than a generic notice. Do not apologise for the decision. Customers respond to clarity.
The redirect offer runs in the final 30 days and should carry a genuine benefit tied to the online channel rather than a discount. Free shipping on the first online order, early access to the next release, or a loyalty point balance carried over from in person purchases all work, and none of them train customers to wait for a sale. Track one number through the transition: the percentage of channel customers who place an online order within 90 days of closure. Anything above a third means the exit worked. Below 10% means the capture step was skipped, and that is worth knowing before the next channel decision comes around.
A channel is profitable when its revenue exceeds every cost that would disappear if the channel closed, measured across twelve months rather than a single season. Include rent or pitch fees, staffing at the rate you actually pay, travel, insurance, permits, damaged or unsold inventory, payment processing, and the opportunity cost of founder hours. Most merchants measure only the first two and reach the wrong conclusion. Once you have the figure, convert it to contribution margin per dollar of revenue and rank every channel against that number. A channel that ranks last on that measure for two consecutive quarters, with no untested lever remaining, is a closure candidate rather than a fix candidate.
Close a pop-up or market channel when it fails at least three of four audit signals: negative fully loaded contribution margin across two quarters, high customer overlap with your existing email list, consumption of founder weekends or a key operator, and no untested lever remaining. High customer overlap is the most commonly missed signal, because a channel that converts people already on your list is relocating purchases rather than acquiring customers. If you have never tested a different location, day, price, or product mix at that channel, run one clean test first. Closing before you have that evidence risks abandoning a channel that had a straightforward fix available.
Revenue usually dips for one to two quarters and then recovers if the freed capital and attention move to a channel with better unit economics. The dip is largest when the closing channel served customers who were genuinely new rather than already reachable online. BOTB provides a documented example of the alternative outcome: revenue for the six months covering its retail exit rose to £7.12 million from £5.54 million, with adjusted profit before tax up 14.7%, because the business used the exit to focus pricing and product on its online customer base. Plan for a dip, capture contact details before closing, and measure recovery at the 90 day mark.
Fix it first when one meaningful lever remains untested and you have the budget to test it, and close it when costs have climbed for three years and every obvious lever is exhausted. The distinction is evidence, not sentiment. A channel with positive contribution margin trailing your best channel by under 20% is a fix candidate. A channel running negative fully loaded margin for two consecutive quarters, staffed by someone you need elsewhere, is not. Founders most often get this wrong in the direction of holding on, because the channel is frequently the one the business started with and closing it feels like disowning the origin story.
Capture contact details before the final trading day, announce the closure directly to those customers with a specific date and reason, then run a redirect offer during the last 30 days. Shopify POS creates customer profiles at checkout, which makes the capture step a configuration decision rather than a manual one. The redirect offer should carry a real benefit tied to the online channel, such as free shipping on a first online order or early access to a release, rather than a discount that trains customers to wait for sales. Track the percentage of channel customers placing an online order within 90 days of closure. Above a third indicates the transition worked.
