The arithmetic of the problem
Grow Fragrance lost $642K last year on $3.4M of net revenue. That's a real loss on a real business — not a one-time write-down, not a single bad bet. It's the shape of a company whose acquisition math has been underwater for longer than we realized, and is now getting worse faster.
Here's what's actually been happening.
Through 2025, Grow spent about $37 to acquire a Shopify customer. That customer's first order generated about $23 in contribution margin — what was left after product, shipping, and platform fees. Subtract one from the other: the business lost about $14 on every new customer at the register. Then about 30% of those customers came back, and the margin from their repeat orders funded the other 70%, the day-one losses, and most of the overhead. Not comfortable — the math was already negative on day one — but it worked, in the sense that the retention engine was doing more lifting than the P&L made visible. The flywheel was subsidizing the funnel. That's not a line item in any chart of accounts, but it was the real story.
Then the inputs moved. Meta's cost per thousand impressions is up about 20% year over year across DTC. Grow's own CAC rose from $37 in 2025 to $41 in Q1 2026 and to roughly $55 over the last 11 weeks. First-order contribution margin did not budge. The day-one loss widened from -$14 to -$18 to about -$32. The business didn't start losing money in April. It's been losing money on acquisition for a while. What's changed is how much it's losing per customer, and how fast that number is moving in the wrong direction.
You can see the pressure in one number. Shopify's share of orders coming from brand-new customers dropped from 51% in Q1 2025 to 42% in Q1 2026. April data is partial, but it's trending materially lower still. The retention engine is now carrying a larger share of the business each month, not by design but by default — there are simply fewer new customers coming in. That's a feature and a warning.
A feature, because it tells you definitively that the product works. Customers who try Grow once and come back a second time are an endorsement of the chemistry, the scent, and the brand. Most failing DTC businesses have the opposite problem — good marketing dressing up a mediocre product. Grow has the better problem: a product that proves itself in the data, attached to a financial structure that has been quietly losing money to find customers for it.
A warning, because a retention engine can only carry so much weight. When the day-one loss was -$14 and the retention 30%, the math was sustainable. At -$32 and rising, it is not, no matter how loyal the customers who come back. The flywheel is real but it has a speed limit.
So the problem statement is simple. Grow is a good product being sold through a financial structure that requires cheap, predictable customer acquisition — and that input has broken, not suddenly, but progressively over a period that is now approaching two years. The turnaround is the work of rebuilding the structure so the product's quality can do its job.
A note on methodology, because this section's numbers corrected two I had wrong before. The CAC figures use Meta plus Google media spend divided by unique first Shopify orders (counting orders, not line items — a small but consequential distinction). The retention analysis is Shopify-only, because Amazon does not expose customer-level data; 18% of net revenue sits outside this view. First-order contribution margin is drawn from the datahub's computed CM column, which is populated for 97% of 2025 first orders but carries a margin of error tied to incomplete BOM coverage on newer SKUs. The direction of that error is downward — meaning true first-order CM might be $2-4 lower than reported — which makes the thesis firmer, not weaker.
The shape of the fix
There are five moves. None of them is clever. Cleverness is what you reach for when a business has no real advantages, and we have several. What we need is execution on the obvious things, in the right order, without getting distracted by the 100 other things that would feel productive.
The diffuser changes the math
If I had to pick one move that makes everything else work, it's launching the nebulizer diffuser as Grow's new anchor product. The reason is arithmetic, not romance.
A first-order contribution margin of $23 can support a CAC up to $23. Beyond that, the business is paying for each new customer rather than earning from them. In 2025, with CAC at $37, the business was already paying $14 per new customer. Today, with CAC at roughly $55, it's paying more like $32. The acquisition funnel has become an outright subsidy, funded by a retention engine that is not structurally designed for that role.
The diffuser rewrites this equation from the other side of the ledger. A first order that includes a $100+ device generates several times the contribution margin of a spray-first order, because absolute dollar margin on a higher-priced product is larger even at similar margin percentages. A diffuser-first customer plausibly delivers $40-50 of first-order contribution margin, possibly more. We'll know more precisely once we've locked the landed cost. At those levels, the business can absorb a $40 CAC and still earn $10-20 on day one. It can acquire through channels that never worked at spray economics. It can tolerate being wrong about something without going broke.
There's a second effect that matters more over time. Every diffuser customer creates a recurring revenue stream — refill oils on a 60-day cycle — that Grow does not currently have in any form. A spray customer is worth the margin on that spray plus whatever repeat behavior follows, which is probabilistic. A diffuser customer on a refill subscription is worth a stream of known cash flows. One is a hope. The other is an asset on the balance sheet that doesn't show up on the balance sheet.
What makes this worth concentrating on is that it isn't a hypothesis. The diffuser is built. It works. The scientific achievement — a nebulizer that puts plant-based fragrance into the air at performance levels that compete with synthetics — is the one thing Grow has that nobody else has. That's the definition of a moat, and it's a moat the company hasn't yet monetized.
The launch happens in Q3. The risk is not product failure. The risk is underfunding, poor positioning, or bad timing against the holiday quarter that matters most. The single number to watch post-launch is the refill subscription attach rate — how many diffuser buyers set up a refill cadence at checkout. That number tells you whether what we have is a product or a franchise.
The price increase is not about margin
Everyone focuses on what a price increase adds to the top line. Raising sprays from $16 to $22 probably adds somewhere between $200K and $400K of annual revenue, depending on elasticity we haven't measured yet. That's real money. But it isn't the most important thing the price increase does.
The most important thing it does is make the acquisition math forgiving. At $16, first-order CM is $23. At $22, first-order CM crosses $30. Even at today's elevated $55 CAC, the gap narrows meaningfully, and for the channels running closer to $40 CAC the business would break even on acquisition at the register rather than waiting for repeat purchases to bail out day one. That's the difference between "hope they come back" and "profitable from the first order." The 30% who come back stop being a required subsidy and become pure upside.
There's a positioning truth at work too. A 5-oz bottle of plant-based fragrance made with ingredients that cost several times what synthetics cost has no business being priced at $3/oz. That's the price of a slightly-premium commodity. The ingredient story supports pricing closer to the tier Juniper Ridge and Aesop occupy — $8-20/oz. We are not going there in one move. We are moving from $3/oz to about $4.40/oz. The ceiling is higher than we have ever priced against.
Two caveats. First, the $200-400K revenue estimate is a planning range, not a forecast. We haven't run a price elasticity test, which is exactly what the 90-day test on three hero scents is for. The test tells us whether the range is right, low, or high. Don't commit a P&L to it before the data comes in. Second, plastic bottles don't hold $22 pricing. Customers who pay premium prices reasonably expect premium packaging. We may need aluminum, which adds $1-2.50 per unit in cost (industry benchmark; actual vendor quotes are next). That's absorbable against a $6 price increase. It's also a commitment — new tooling, new photography, inventory transition on the plastic. Decide in 30 days.
The acquisition function needs a rebuild, not a diet
The most misunderstood move in the program is the Meta spending reduction. The original plan called for cutting 25-30% of paid-media spend, which sounds like austerity. It isn't. It's the recognition that the marginal Meta dollar has stopped earning its keep. Cutting bad spend isn't discipline for its own sake — it's freeing up capital to deploy somewhere better.
What has changed since the first version of this plan is that the acquisition problem looks more structural than it did in April. New customer acquisition on paid channels is down roughly 48% year over year over the last 11 weeks. CAC in the same window rose from $38 to $55. Some of that is real macro — Meta really is more expensive for everyone, roughly 20% more. But not all of it, and probably not most of it. A meaningful share is the agency's execution, creative that has gotten tired, and a campaign portfolio that spends aggressively on retargeting existing customers who would have bought anyway.
The right read isn't to cut Meta by a percentage. It's to rebuild the function. New-customer ROAS becomes the primary measurement instead of blended MER. Branded search — the part of Google spend that pays to show up for customers who already know us — gets capped, because it cannibalizes organic traffic. Performance Max on Google, which is the right tool for a catalog-driven business and has been underfunded, gets expanded. Amazon advertising, which has been starved despite generating strong returns on spend, gets moved back to its budget floor. Creative refresh gets a cadence: every 30-45 days on top-spending campaigns, not whenever somebody remembers.
A note on scope. When the plan says "ad spend" or "paid media," it means the $1.1M the business spent on Meta, Google, Amazon, Collabs, and Pinterest in 2025. The P&L's Total Marketing line is closer to $2M for the same year. The difference — agency retainers, creative production, tool subscriptions, affiliate, and other marketing labor — isn't captured in our media tables and is a separate workstream for discipline, not the one this section addresses.
Whether the agency doing this work is our current one, changes, or whether we take some of it in-house, is a conversation happening in parallel. The function-level changes are the thing. The people doing them are secondary.
The retention engine is already working
There's one part of the business that's functioning extraordinarily well, and it's the one we've invested in least. Grow's email program, through Klaviyo, was a touchpoint on roughly $170K of attributed revenue in Q1 alone. That number is an attribution total — it overlaps with paid channels, because a customer who clicks a Meta ad and then clicks a follow-up email shows up in both sets. The incremental revenue — what wouldn't have happened without email — is somewhere in the $70-120K range for the quarter, per typical DTC attribution patterns. That's still meaningful, because the variable cost of email is essentially zero. It's the closest thing we have to free money.
And it's operating below its potential. The Welcome Flow — the automated sequence that greets a new subscriber — earns $1.23 in revenue per recipient. That is excellent. The Post-Purchase flow that's supposed to turn a first-time buyer into a second-time buyer earns $0.06 per recipient. That's a 20x gap on the single conversion that matters most. The median time between a Grow customer's first order and second order is 85 days, which tells you the current timing of the follow-up emails is almost certainly wrong. The winback flow, for customers who haven't bought in a while, is similarly underperforming.
Here's the shape of the work. Katelyn owns the content. I own the plumbing. The ambition is to take the share of revenue coming from repeat customers from 54% last year to 62% by the end of this year. On a target of about $2.9M of Shopify revenue in 2026, that's $230K of incremental retained revenue we already know how to generate, because we already do it well on the parts of the program that are working.
There's also a set of tools we haven't really used yet, which fall under the broad heading of behavioral economics. Loss framing — "your scent quiz result expires in 48 hours" — instead of gain framing. Endowment — "the scent you picked is reserved for you" — instead of description. Default bias at the refill subscription checkout, where the subscription option is pre-selected and customers opt out rather than opt in. These aren't clever tricks. They're the basic grammar of how consumers actually make decisions, and they are free to implement.
We'll run A/B tests on coupons on the 1st-to-2nd order flow specifically, because that's where the conversion cliff is steepest and the upside is largest. Not because coupons are a strategy, but because finding the right offer, at the right time, for the right customer is worth a lot of money in a business where retention already works.
A scope note here too: this retention work is Shopify-only. Amazon, which is 18% of net revenue, does not expose customer-level data, so we cannot run the same kind of flow analysis on Amazon buyers. Amazon retention is a separate question, answered through Amazon Brand Analytics, and isn't part of the engine we're describing. The full operational breakout for this workstream lives in the Email & Retention Project.
Wholesale is the quiet channel
The loudest parts of the turnaround are the diffuser and the price increase. The quietest part may prove the most durable, which is pushing wholesale from $59K a year to $300K or more.
Wholesale has several virtues DTC does not. There's no CAC, because retailers find us or Whitney goes out and finds them. The orders are larger — the average Faire order last quarter was $113, almost double what it was a year earlier. The retailer does the storytelling; we don't have to fund a creative refresh every six weeks. And the diffuser is the ideal wholesale product: a $100+ device with a refill revenue stream is exactly what premium retailers want to carry.
Whitney is taking this on at zero incremental payroll. The target is 150 accounts by year-end, covering yoga studios, premium grocery, boutique retail, and boutique hotels. Two trade shows, Faire and Handshake for discovery. MAP pricing enforced so retail margins work when our own DTC pricing rises.
This isn't the part of the plan that excites people. It's also the part of the plan that is hardest to screw up and most likely to still be generating revenue five years from now.
Discipline is what gets held flat
The last move takes no effort and requires the most willpower, which is to hold overhead at roughly $1M while revenue grows toward $4.5-5M. Every business in a turnaround looks at a rising top line and finds new reasons to hire, new tools to license, new initiatives to fund. Most of those reasons are real. Almost none of them are worth what they cost.
There's a parallel workstream of things on the company's plate that aren't turnaround moves but could become turnaround problems. Warehouse staffing. Operating expense management. The data platform. The MRP rollout. The real estate decisions. Each has to execute without blowing the overhead line. If any one of them needs $100K+ of incremental spend, it comes up for re-scoping before it gets approved.
The tail of the product catalog gets cut too. Keep the top 20% of SKUs that drive most of the revenue, retire roughly 25% of the tail. This simplifies everything — forecasting, warehouse space, working capital, the photography pipeline, the email calendar. This is the part of the plan where the instinct to hold onto every launch has to be resisted. Not every scent is a winner. The ones that aren't need to stop consuming the oxygen of the ones that are.
How to read the math
The five moves, executed, take the business from $3.4M in net revenue and a $642K loss to somewhere between $4.2M and $4.8M of net revenue and net income between +$100K and +$400K.
The range is wide for a reason. The diffuser is the swing variable. If it launches well and the refill subscription takes hold, the business is on a different trajectory entirely — not because the diffuser itself is huge in year one, but because every diffuser customer creates a recurring revenue stream that compounds. If it launches poorly, or the positioning is wrong, the low end of the range is where we land. The low end still gets the business to break-even, which is the floor worth defending. Break-even is what turns Grow from a company that needs more capital into a company that can fund its own next chapter. Profit is optional. Not losing money is not.
Three things to understand about how soft this range is. A diffuser that comes in at half of target ($150K rather than $300-500K of Year 1 revenue) costs about $100K of the bottom-line range — a real hit, but not fatal, because retention and price increases carry their own load. A price-elasticity outcome worse than the planning assumption (say, 30% volume loss rather than 15%) compresses the spray-revenue contribution to closer to $100K than $300K. And a failure to restart new-customer acquisition — if CAC stays at $55 and first-order share keeps sliding — widens day-one losses by more than the other moves can absorb, and pushes us toward the bottom of the range even if everything else hits plan. The moves compound upward when they work together. They can compound downward too.
What would tell us it's working
By July, the first-order share of Shopify revenue should be recovering from its April low. If the acquisition function restructure is working, that number trends from mid-20s back toward 35-40%. If it isn't, we know early, and we don't wait to adjust.
By the end of Q3, the diffuser should have shipped meaningful volume. Not a breakout — a clean launch. The number to watch isn't total diffuser revenue but refill attach rate: the percentage of diffuser buyers who set up a subscription at checkout. Above 40% is a flywheel. Below 20% is a product that needs a merchandising rethink. I don't have category benchmarks on this specific number — it's my estimate of what healthy looks like, and we should replace it with actual DTC subscription comparables before launch.
By the end of Q4, the email program's 1st-to-2nd order conversion should have measurably improved as a result of the flow audit and the timing changes. The behavioral-economics tests should have at least one clear winner. The aggressive 62% repeat-revenue goal does not have to hit — 58% would be evidence that the program is moving.
By January 2027, the wholesale door count should be in three digits. Overhead should still be at about $1M. The P&L should show the shape of break-even, even if the full-year number isn't there yet.
If those things are true, we compound. If they aren't, we reassess with the same honesty we're trying to bring to this plan. Not with a new five-year roadmap.
The last word
There's a version of this business in which Grow is a durable premium fragrance brand, selling a defensible product at defensible prices through a healthy mix of DTC, wholesale, and marketplace channels, with an email program that compounds customer value and a diffuser franchise that didn't exist in 2025. That version is one year of disciplined execution away.
There's also a version in which the company tries to fix too many things at once, spends the next 12 months writing plans instead of shipping products, and runs out of runway while building the machinery to run a much larger company than it has. We've done more of the second than the first.
The point of this plan is to stop doing that. Five moves. A 12-month clock. The math we already have, corrected where we needed to correct it.
A last discipline worth naming. The numbers in this plan are better than the numbers in the plan I gave you two days ago, because we stepped back and checked them. Two of them turned out to be wrong in ways that would have mattered. That discipline — regenerating, re-pulling, red-teaming a thing before it goes to the team — is the reason the plan is more defensible now than it would have been. It's a habit worth keeping as we execute. Nothing we put in front of Katelyn, Nikita, or anyone else should be trusted just because it was produced confidently. Including this document.
The mission — real fragrance, made with real plants, that performs as well as anything made with chemicals — survives if the business does. The business is asking us to be disciplined on its behalf.