The Usual Suspects
Margin leaks don't announce themselves. Nine ways DTC brands lose margin every day — and the defense for each.
Every function in your company watches revenue. Your agency reports ROAS. Your ops team ships orders. Your dashboard celebrates sales. Margin — the thing you actually keep — has no owner. It erodes quietly, in places nobody is paid to look, and shows up months later as "we did more revenue and made less money."
Gotama is the hire that owns it. Every day, he audits your store's economics line by line and puts a signal on the board when he finds a leak: what's happening, what it's estimated to cost, and what to do about it.
28-day free trial · Cancel anytime
- Attack vector:
The ROAS Mirage
Revenue metrics can't see costs. Discounts deepen, shipping rates creep, fees stack — and ROAS doesn't move, because ROAS was never measuring profit. You can hit your target every week while your blended efficiency quietly sinks.
The defense —Efficiency Break DetectedMAC 5Gotama tracks your true marketing efficiency — actual Shopify revenue against actual ad spend, so a broken pixel can't fake it — versus your own four-week norm. He fires only when the break holds on consecutive checks, puts a per-day cost on it, and names the driver: order value falling, or acquisition getting pricier. In the POAS Lab, every product is graded SCALE, HOLD, or CUT on profit-on-ad-spend, not revenue.
- Attack vector:
The Algorithm Off the Leash
PMax, Advantage+, and their cousins spend on their own judgment. Most days that's fine. Some days one of them locks onto the wrong audience or a broken landing page and triples its budget before anyone checks the ad account.
The defense —Rogue Algorithmic SpikeMAC 4Fires when an algorithmic campaign runs at multiples of its own clean baseline two days running while your store's revenue doesn't follow. When your scaling is working, he stays silent — he flags the runaway, not the win.
- Attack vector:
Zombie Spend
Ads fatigue. Audiences saturate. Creative dies. But dead ads keep billing, and in a busy account nobody notices the campaign that stopped earning three weeks ago.
The defense —Wasteful Ad DetectedMAC 1Flags ads still spending after their returns have collapsed, with the estimated daily bleed on the card.
- Attack vector:
Flying Blind
A theme update, a new app, a checkout change — and your pixel or conversion API quietly breaks. The ad platforms keep spending; they're just optimizing on noise. Most brands find out weeks later, from a hunch.
The defense —Tracking Anomaly DetectedMAC 3Spots when a platform's reported results decouple from what your store actually sold — the signature of a broken feed — so you fix it in days, not quarters.
- Attack vector:
The Code That Got Away
An influencer's code hits a deal forum. A staff discount escapes the office. A retired code resurfaces on a coupon site and starts stacking. Every redemption still looks like a sale.
The defense —Discount Code AnomalyOAC 2Four tripwires — a new code going viral, a code appearing in channels you don't own, discount depth spiking, penetration surging. Tell Gotama your planned launch prefixes; he watches everything else.
- Attack vector:
Selling Below Cost
A supplier raises costs. A price gets set by feel. A discount stacks where it shouldn't. Some SKUs quietly go negative — and because they "convert," your algorithms scale exactly those. Your best seller becomes your biggest loser.
The defense —Negative Margin SaleOAC 4Line-level math on every order, every day — price, discount, tax, landed cost — flagging SKUs selling below cost at material volume, with the actual orders attached as evidence.
- Attack vector:
The Quiet Return Problem
A bad batch. A size that runs small. A product page that overpromises. Blended return rates hide which SKU is responsible while it claws back the margin you thought you'd earned.
The defense —Return Rate SpikeOAC 5Per-SKU return pressure against that SKU's own history, damaged-versus-restockable split, and the estimated margin leak — so you find the batch, the size curve, or the PDP, not just a scary average.
- Attack vector:
Paying to Promote Empty Shelves
Your ads don't know your inventory. When the variants that actually convert run low, the campaigns keep spending — buying traffic for sizes and colors you can't sell.
The defense —Dead-End Ad SpendOAC 3Watches days-of-cover on every variant, weighted by how much revenue each one drives, and tells you to cut the ads when a product's converting variants are days from empty — before the spend is wasted, not after.
- Attack vector:
The Settings Leak
Tax configuration is set once and audited never. Gift cards taxed at purchase instead of redemption. International orders shipping with no tax charged — VAT you'll owe either way, paid out of your margin.
The defense —Tax Configuration RiskOAC 1Order-level tax checks that catch the misconfiguration while it's a handful of orders, then verify the fix against the next real transactions.
Defense funds offense.
When wasteful spend gets cut, Scale Opportunity Found MAC 2 flags the products with safe, profitable headroom to take that budget — same evidence standard, opposite direction.
Every signal closes its loop.
Each card carries an estimated impact, a prescribed action, and an owner on your team. When they mark it resolved, Gotama verifies over the following one to two weeks that the fix actually held — and tallies the recovered margin every month, so defending margin becomes something your team can see itself getting paid for. And he knows the difference between fixed, paused, and new normal — pausing your spend defers a card rather than closing it, and a shock that stabilises into a new baseline resolves honestly instead of burning red forever.
Tuned for real stores, not lab conditions.
Gotama knows a planned sale from a leak, a deliberate scale-up from a runaway, and a seasonal pattern from a problem. Signals are floored to materiality — he won't nag you about pennies.
28-day free trial · Cancel anytime