Agentic analytics for multi-store retailers and brands
Gross margin across independent brands is at a record high.Net margin is at an all-time low.
Across 300 independent store owners with $3.5bn in combined revenue, 2026 gross margin was 49.5% and net margin was 10.6%. Most of the 39 points in between are spread across your stores, storefronts and collections, and most groups see them a month late, one store at a time.
Left on the table each year
For a group with 30 stores and $2.4m in ad spend that sees margin by store and storefront a month late instead of every week.
30 minutes with someone who has connected Shopify, Klaviyo, ERPs and phone systems before. Or run your own numbers first.
of multi-store retailers have revenue coming in through a channel nothing in their stack can see: phone, wholesale, showroom or ERP.
of apparel and sporting brands can’t tell which campaign, creative, coupon or collection drove their revenue.
of multi-store retailers rely on spreadsheets, forms and manual math to keep reporting running.
From in-depth research with 19 multi-store retailers, 31 home and lifestyle retailers, and 15 apparel, footwear and sporting brands. In 10 of the 19 multi-store retailers, the owner or managing director joined the call.
Ad spend connects in a click. Revenue often doesn’t.
Your order system, ERP, phone lines, wholesale ledger, marketplace and regional storefronts don’t plug in that easily, and that’s where much of the money lands. So the spreadsheet survives, even in companies that already bought a BI tool.
The number you need is current margin for each store, storefront and collection, next to the spend behind it. Every system reports revenue, and blended ROAS only measures the share of revenue the platforms can see.
Which stores and storefronts carry the group
A group average hides the spread, and the spread is where the profit is. HFA’s benchmarks put close rate at 34% against 38% for the top tier, and average ticket at $1,251 against $2,394. Brands see the same thing across regions: sell in AU, NZ, UK and EU and the platform hands you one blended return, while one region props up another and one collection loses money on every order it ships.
A store or storefront that has been slipping since spring looks fine inside a monthly roll-up.
What we build for it
An AI agent ranks every store, storefront and collection each week on margin against its own baseline, and names the ones that moved, with the reason: ticket size, close rate, returns, product mix, or spend that went to the wrong place.
Proof
Three businesses that were doing this by hand.
Here’s what each of them was dealing with.
Every day, one person typed the numbers for 56 showrooms into Excel. What they wanted to know was how all 56 compared that morning.
- Daily, by hand, every store
- One person holding the whole picture
- More work with every new showroom
Qasr Alawani
They wrote 40 SQL queries by hand to see their bestsellers, and their reporting only kept 40 days of history. Production was being planned on a window too short to plan from.
Donsje
“Is not acceptable… and that’s why we called you, frankly.”
What prospective private equity investors told the founder about running the business off a Google Sheets dashboard.
Founder, Scent Fill
Your numbers
What is the gap between your stores worth?
Enter two numbers. We fill in the rest from the retailers and brands we studied.
We assume about 7 in every 30 stores run $250k a year below your median, and numbers reach you monthly.
Left on the table each year
That’s $4,178 for every week you wait.
Ad budget moved to the channels that actually drive sales, online and off
Revenue won back from weaker stores because you spot them sooner
Staff time back from typing store numbers into spreadsheets
How we worked this out
35% of revenue comes through channels the ad platforms can’t see (10 of 19 retailers named the system). With monthly, store-by-store reporting, about a third of the spend problem and a fifth of the store gap run before anyone notices, and we count a quarter of each as recoverable in the next cycle. Re-keying is 15 minutes per store per day, 250 days a year, at $30 an hour, like the 56-showroom case above. We only count what you could realistically have caught, and nothing from returns, tariffs or collection mix.
Bring your store list. You’ll leave knowing which of your systems we can read, what it would take, and what it costs.
What a collection makes once it lands
Average US tariffs on apparel imports rose from 14.7% in January 2025 to 35.1% a year later, and brands absorbed 58% of that cost. Section 232 put 25% on imported upholstered furniture, cabinets and vanities, with increases delayed rather than cancelled. Returns run a quarter to two fifths of orders in apparel, and fewer than half resell at full price.
All of it lands on the same question: what does a collection or SKU actually make once everything has happened to it? Almost nobody can answer that from the systems they have.
What we build for it
An AI agent tracks landed cost, returns and sell-through against the buy for each collection and SKU, reruns it when a duty rate or freight cost changes, and flags the lines that stop making money before you place the next order.
Numbers that match the source
Numbers that don’t match the source system are the one problem we saw cost brands real money. It caused a paused subscription, sent one brand back to checking revenue by hand in its store back end, and was behind the only outright loss in the apparel group we studied. Every number we produce links back to the record it came from, and you can open it. If it doesn’t match Shopify or your ERP, we don’t ship it.
Before you ask
Questions we usually get before the call.
“Our ERP or Shopify already reports this.”
Each reports what’s inside it, one store at a time. Your ad spend, phone system, marketplace, wholesale and finance ledger live elsewhere, and margin by store needs all of them in the same row. Joining them is the work, and it’s why the spreadsheet is still there.
“We already run Triple Whale or Northbeam.”
Those tell you what your paid media returned. This shows what a store, storefront or collection makes after landed cost and returns, including the revenue that never touches an ad platform.
“We could export it and ask an AI.”
Chat tools cut off large exports without telling you, give a different answer next week, and do nothing on Monday unless someone asks again. What you need is the connection to your systems and a report that turns up on its own.