Everything about your business, in one place. Five layers making sense of it.
Clariyo is an intelligence system for your store. It holds your business context, reasons across every function, surfaces proven actions to drive growth, and earns your trust with honesty & confidence at every step.
!Should I scale this campaign, or is it already underwater once returns are in?
✓Which acquisition channel brings buyers who actually come back?
!Which of my bestsellers are actually losing money after ads and returns?
✓Why did margin drop 6 points this week?
!Where is my cash right now?
✓Is my biggest seller also my most profitable?
Clarity, and what's underneath it.
Behind every answer is a stack of intelligence, not a lookup. It reads the data, holds the context, runs the investigation, drafts the action, and earns the trust — layer by layer. Step through to see how.
ShopifyKlaviyoMeta AdsStripe
Why did email revenue drop last week?
Joining Klaviyo sends with Shopify orders across 14 months...
SELECT flow_id, sends, attributed_revenue FROM klaviyo_sends JOIN orders USING (email)...
Attributed revenue by week
1,204 rowsPin to briefingOpen an investigation
Data
01 / 05
Connects your stack and captures everything, read-only.
Known failure patterns
⊙
Attribution drift
Ad platform and Shopify orders diverge — clariyo reconciles both.
✉
Silent flow failures
A flow stops sending; clariyo ties the revenue dip back to it.
□
Stockouts throttling ads
Spend runs on a SKU that can't ship — clariyo flags the mismatch.
⇄
Repeat-rate decay
New-customer growth can mask a falling repeat rate underneath.
Context
02 / 05
Starts from your store's likely causes, not a blank query.
Investigating — sales dip
~2 min left
why are my sales down last month
Thinking — this reads as an investigation, not a quick lookup.
Revenue down 13.7% Aug vs Jul. Validating the premise before I plan further.
Investigation plan
✓
Validate the premise — revenue actually down
✓
Decompose revenue — orders vs AOV
Rule out returns and line-type shifts
Running now
Synthesize root cause and draft the fix
Subagents active
● revenue_x_aov — done, orders explain ~96% of the drop
Investigate
03 / 05
Opens automatically. Tests hypotheses against your real data.
Needs your decision2
Marketing
Cut spend on 3 negative-margin SKUs
From the margin drop investigation
Mark done
Finance
Open a claim for SAR 34k of unsettled orders
Courier shortfall, 6 orders over 30 days
Draft claim
In play2
Ops
Switched courier for orders over 4kg
Implemented 9 days ago
● Measuring · ~2 weeks
Marketing
Added BNPL surcharge at $40 threshold
Live 3 days
● Measuring · ~10 days
Measured2
Ops
Tightened the delivery promise
Returns down 22%
Est.+$2.2k/mo
Actual+$2.8k/mo — beat
Marketing
Paused Meta prospecting set
Contribution margin improved
Est.+$1.4k/mo
Actual+$1.6k/mo — beat
Action
04 / 05
Drafts the fix. You approve. It never sends on its own.
34
Actions proposed to date
$1.8M
Confirmed value delivered
21
Closed calls, visible either way
Confirmed
Reduced checkout latency after Q2 traffic surge
+$142K GMV · High confidence
Confirmed
Reallocated spend from underperforming channel
+$61K margin · High confidence
Did not hold
Bundled discount test on slow-moving SKUs
No measurable lift · Low confidence
Net impact this quarter
+$203K
Trust
05 / 05
Confidence levels, full trace, no fix left untracked.
All your sources connected, and available.
Ads, Shopify, email, support, spreadsheets — every source lands in one place, read-only. Ask in plain English and get an answer back immediately. No SQL, no analyst in the loop, no dashboard to build first.
It already knows how DTC stores break.
Clariyo is trained on the failure patterns of ecommerce operations — attribution drift after an iOS update, a flow that silently stops sending, inventory that quietly throttles a winning ad set. It starts from the likely causes instead of guessing, and gets sharper on your store every week.
An agent for your every "why is this off?"
Clariyo watches your core metrics across all sources. When one breaks trend, it opens an investigation, forms competing hypotheses, and tests each one against your real data — ads, sessions, orders, email — until the cause holds up.
The fix, drafted and ready to ship.
Once Clariyo proves a cause, it drafts the fix — a paused flow to re-enable, a bid to pull back, a segment to exclude — ready in the tool you already use. Nothing sends itself. You review the draft, approve it, and Clariyo tracks what happens after.
It shows its work.
Every fix Clariyo proposes is tracked against real revenue — the wins and the misses, both counted. It gives you a confidence level up front, and says plainly when it can't prove something. Every answer comes with the trace behind it, so you can check the math yourself.
What it watches, every night
Eleven dimensions of the business, held in one engine.
Every dimension checked against the other ten. A returns problem in one city, a delivery promise and three products on ad spend surface as one story, not three.
One engine
11 dimensions
Financial health
Contribution, margin trend, cost structure.
Demand & traffic
Sessions, sources, seasonality against your own history.
Conversion
Step by step, by device, city and channel.
Products
What moves, what earns, what quietly loses.
Marketing efficiency
Reconciled return, fatigue, wasted spend.
Retention
Cohorts, repeat behaviour, churn risk.
Inventory
Cover, stock-out risk, cost of being down.
Fulfilment
Delivery promise, courier performance, refusals.
Customer experience
Returns reasons, complaints, repeat friction.
Payments
Fees, settlement timing, failures.
Storefront reliability
Tracking breaks, checkout errors, speed.
ONE CLOSED LOOP
It works your business whether you open it or not.
A dashboard
Waits to be opened, and shows every store the same view.
clariyo
Went through your business last night, found most of it healthy, and came to you about the three that were not.
The Monday briefing. What you earned, the three things that need a decision, and the eleven checks that came back clean.
Monitor
Every source monitored against your own history, not a benchmark.
Investigate
A planned analysis across marketing, margin, customers and cash at once.
Conclude
Cause first, with evidence, confidence, and what was ruled out.
Strategise
The course of action, modelled, with the money attached and a draft ready.
Verify & track
You mark what you did. It measures the result against a control.
Learn
Across every step. Results and your corrections stay in its memory.
HOW IT WORKS
It brings you the problem before you have noticed it.
Everything gets checked. Only the most prioritized makes it to you. Here is what happens to one night's worth of signal.
300+
Metrics monitored
Across the eleven dimensions, every night, against your own history.
14
Moved outside range
Most of these are noise, seasonality or a one-off.
6
Worth a look
Persistent, and commercially material once sized in money.
3
Escalated to you
Ranked by what is at stake, not by how unusual they look.
1
Taken to root cause
Investigated overnight, before you opened the app.
03:12Contribution margin down 6 points week on week. No one asked. It sized the exposure at USD 5,400 a month, ranked it first of the three, and went looking for the cause.
Ruled out
✕
Product mix
Explains 0.3 of the 6 points.
✕
A fee change
No gateway or courier rate moved.
✕
Discount timing
Depth flat at 0.8% of gross.
✕
Supplier quality
Same batch sells clean elsewhere.
✓ Confirmed
USD 5,400/mo
Returns in one city
Tripled in Kansas after the delivery promise slipped to 5 days.
AND THE PART IMPORTANT TO JOIN
Snap scaled those same three products by 40% the same week, so you were paying to sell more of the thing that was coming back. Three actions, USD 5,400 a month.
+5
It also found 5 opportunities worth acting on and held them back; they were worth less than these three.
Why you can trust it with a real decision
It tells you what it could not prove.
Confidence
Every conclusion carries a confidence level and the explanations it tested and ruled out.
Coverage
If a connector is broken, it says so and stops. It marks every estimated figure.
Silence
Each morning: how many checks ran and how many came back clean. Proof of work, not noise.
Hit rate
Every recommendation is measured against a control. The misses sit on the record beside the wins.
Every system you run, under one roof.
clariyo connects to your storefront, ads, payments and lifecycle tools, read-only, and reasons across all of them together.
