Daily Ops overview

Audience: Merchant·Plan: starter·Read: 8 min·Updated: 2026-08-03

In 30 seconds

• One worklist for the whole platform. Ten engines stop shouting in ten places and emit into a single ranked queue.
• Ranked by money, not by noise — (amount at stake × confidence) ÷ effort. A cheap, high-confidence, high-value fix beats a vague hunch every time.
• Executive Pulse puts revenue driven, recoverable now, agent-ready %, open actions, and brand authority on one line.
• Five tabs: Actions · Weekly Plan · Revenue Leaks · AI Readiness · Pages.
• Findings auto-resolve when the underlying problem goes away — the list shrinks by itself.

In plain English

Your shop has a lot of moving parts, and each part notices different problems. One watches your prices, one watches your stock, one watches what customers say in reviews, and so on.

Normally you'd have to check each of them separately, and most people don't have time. So problems sit unnoticed.

Daily Ops is one screen that collects everything they found and puts it in order — most valuable and easiest to fix at the top. You start at the top, work down as far as you have time for, and you can be confident you did the most useful things available.

At the top of the screen are five numbers that tell you how the business is doing right now. Below that are five tabs, each answering a different question.

How to use it

1. Click "Daily Ops" in the left-hand menu.
2. Look at the five numbers across the top. That's your 10-second health check.
3. Work down the Actions list from the top. It's already in the best order — you don't need to decide what matters most.
4. Click any row to go to the screen where you fix that problem.
5. Click "Mark done" when you've handled it, or "Dismiss" if it isn't worth doing. Either way it leaves the list.
6. On Mondays, open the Weekly Plan tab instead — it shows just five things to focus on for the week.

⏱ ~10 min/day · 💳 Starter+ · 🎯 Never wonder what to work on next

Why this matters for your business

The problem with a platform this large isn't that it finds too little. It's that it finds too much, in too many places. Your SEO engine has opinions. Your pricing guard has opinions. Cart recovery, catalog health, review sentiment, replenishment — every one of them is right, and every one of them wants a different tab. So you check three of the ten, the other seven quietly accumulate findings nobody reads, and the single most valuable thing you could have done this week sits on page four of a screen you last opened in March.

Daily Ops removes the choosing. Every engine writes into one shared envelope — the same shape, the same status vocabulary, the same money field — and Daily Ops ranks the combined stream by what it's worth against what it costs you to fix. The top of that list is, mathematically, the best use of the next hour you have.

The second thing it removes is the guilt-driven backlog. Findings here are not tickets you have to close. They expire. A pricing finding that nobody actioned in 7 days resolves itself, because a 7-day-old pricing signal is no longer true. What's on the list is what's live, which is why the list is worth reading at all.

What this typically unlocks

Outcome	Typical result
Surfaces checked per morning	1 instead of 6–10
Time to "what should I do today?"	~10 seconds (read the Pulse)
Highest-value action found	Top of list, not page 4 of an engine tab
Findings that expire unread	Auto-resolved — no manual backlog grooming
Recoverable money made visible	Continuously, with honest recovery rates applied
Team coordination overhead	−70% — one queue, audited actor on every action

What you actually get

Five tabs, each answering a different question. Tab state lives in the URL (?tab=), so any view is bookmarkable and shareable with a teammate.

Tab	Question it answers	Deep dive
Actions	What needs doing, ranked by value?	Actions worklist
Weekly Plan	What are the five things that matter this week?	Weekly Plan
Revenue Leaks	Where is money leaking right now?	Revenue Leaks
AI Readiness	Can ChatGPT/Perplexity actually sell my products?	AI Readiness
Pages	Which pages earn search traffic?	Pages performance

The Executive Pulse

Above the tabs sits a five-metric header. It is composed fresh on every load from five independent engines, and it is deliberately not a dashboard — it's five numbers you can read without interpreting.

Metric	What it means	Source
Revenue driven · 30d	Revenue this platform is attributed for, last 30 days. Sub-line shows your ROI as a multiple of plan cost	Unified attribution ledger
Recoverable now	Honest recoverable total across all open leaks	Revenue leak intelligence
Agent-ready products	% of catalog an AI shopping agent could actually sell. Sub-line shows average readiness score	Agent readiness scoring
Open actions	Count of actionable findings. Sub-line shows how many are critical	Action Center rollup
Brand authority	Composite brand-strength score with band	Brand authority engine

Two design decisions worth knowing:

• Money metrics never get summed. "Revenue driven" and "recoverable" can be in different currencies and mean categorically different things — one is realized, one is a projection. The Pulse shows them side by side and never adds them.
• The headline sentence is deterministic. No LLM writes it. The same inputs always produce the same sentence, so a number that changes means the business changed, not that a model phrased it differently today.

Each panel degrades independently. If the brand-authority engine is unavailable, that tile shows blank and everything else renders — the page never fails as a whole because one reader failed.

How it powers every part of your store

Ten domains feed the single queue. Each is emitted by the engine that owns it, and each routes to the page where you actually fix it.

Domain	What it surfaces	Emitted by	Fix page
leak	Abandoned carts, OOS losses, refunds, failed payments	Revenue leak intelligence	Cart Recovery
agentic_readiness	Products AI agents can't represent or sell	Agent readiness scorer	AI Readiness tab
ai_visibility	Discoverability gaps on AI surfaces	Sales-engine discoverability	Growth Intelligence
voice	Themes from reviews, tickets, and survey text	Customer Voice engine	Customer Voice
pricing	Margin erosion and price-guard breaches	Margin guard	Auto-Price
category	Category and competitive-position shifts	Discoverability scorer	Competitor Analysis
content	Decaying content that needs a refresh	Content decay detector	Content Studio
retention	Replenishment windows and churn risk	Replenishment worker	Second Purchase
merchandising	Assortment and collection problems	Sales engine	Growth Intelligence
executive	Cross-engine strategic findings	Decision Intelligence	Decision Intelligence

Product-scoped findings deep-link straight to that product's editor rather than to a domain page — so "fix this product" is one click from the list, not a search.

How it works (without the technical bits)

The ranking model

Every finding gets a priority score:

score = (value at stake × confidence) ÷ effort

• Value at stake is the recoverable amount when the finding has one. When it doesn't — an AI-visibility drop has no direct currency figure — a notional value stands in by severity, so non-economic findings still rank sensibly against small money amounts rather than sinking to the bottom.

  Severity	Notional value
  Critical	1,000
  Warning	250
  Info	50

• Confidence is 0–1, set by the detecting engine. A statistical certainty scores higher than a heuristic guess.

• Effort divides the score, so cheap fixes float up:

  Effort tier	Divisor	Meaning
  One-click	1	A wired fix exists — press the button
  Quick	2	A few minutes of manual work
  Manual	4	Real investigation required

Ties break on severity, then on most-recently detected.

The practical effect: a $2,000 abandoned-cart leak with a one-click recovery send outranks a $6,000 finding that needs a week of investigation — because you can actually finish the first one today.

Deduplication — one finding per real problem

Several engines can notice the same underlying problem in the same run. Each finding carries an identity key built from domain : kind : scope : scope-id, and duplicates collapse before anything is written.

When two findings share a key, the survivor is chosen in order: higher severity wins, then higher score, then more recent. So re-detection sharpens a finding rather than duplicating it — the row you dismissed yesterday doesn't come back as a twin tomorrow.

Lifecycle — why the list shrinks on its own

Status	Meaning	Counts as actionable?
Open	Newly detected, needs attention	Yes
Snoozed	Deferred, still on the books	Yes
Actioned	You marked it done	No
Dismissed	You judged it not worth doing	No
Resolved	Auto-closed — the signal expired or the problem went away	No

Only Open and Snoozed appear in the worklist and count toward the money rollup, because only those are recoverable now.

Auto-resolution runs per domain on its own clock, tuned to how fast that kind of signal goes stale:

Domain	Auto-resolves after	Why that window
Pricing	7 days	Competitor prices move weekly; a stale price signal is wrong
Agent readiness	7 days	Re-scored on every catalog sync
Customer Voice	21 days	Sentiment themes are slower-moving
Retention	21 days	Replenishment windows are measured in weeks
Category	21 days	Category shifts are structural, not daily
Default	14 days	Balanced fallback for other domains

Leak findings don't use a timer — they resolve on fact. An abandoned-cart leak closes when that cart converts or expires; an out-of-stock leak closes when stock comes back.

Who can see and act

Capability	Permission required
View the worklist, Pulse, and all tabs	growth:read:all
Mark done, dismiss, or reopen a finding	growth:manage:all

Solo merchants bypass permission checks automatically — you don't configure roles to use your own store. Multi-user shops enforce both. Every status change records the acting user's email; on background or automated transitions a stable system identity is recorded instead, so the audit trail never has a blank actor.

All reads and writes are scoped to your shop at the query level, so a finding can only ever be read or mutated by the store that owns it.

Real merchant scenarios

Scenario A — The Monday that used to take two hours

Setup. Home-goods brand, 4,000 SKUs, two-person marketing team. Monday morning previously meant opening SEO, cart recovery, catalog health, reviews, and pricing in five tabs and building a to-do list by hand.

With Daily Ops. The Pulse reads:

Revenue driven · 30d    $184,200   (12× your plan)
Recoverable now         $11,480
Agent-ready products    62%        (avg 71/100 readiness)
Open actions            34         (3 critical)
Brand authority         68         Strong

The top three actions, already ranked:

#	Finding	At stake	Effort
1	47 abandoned checkouts eligible for recovery send	$2,340	One-click
2	"Linen throw — sand" missing price and image	$890	Quick
3	Margin breach on 6 SKUs after supplier cost rise	$1,610	Quick

Result. Monday triage went from ~2 hours to ~15 minutes, and the recovery send at #1 — which had been sitting unnoticed in the cart-recovery tab for nine days — went out that morning.

Scenario B — The finding that ranked above a bigger number

Setup. Apparel brand. Two findings in the same week:

• A $6,200 funnel-drop-off leak. Low confidence (0.3), manual effort — someone has to sessionise and investigate.
• A $1,900 out-of-stock leak on three best-sellers. High confidence (0.9), quick effort — reorder.

Ranking. Scores work out to roughly 465 for the funnel finding and 855 for the OOS one, so the smaller number ranks first.

Why that's right. The OOS reorder was placed in 20 minutes and recovered most of the $1,900. The funnel investigation, started the following week, took three days and ended in "inconclusive." Ranking by raw money would have spent the better part of a week on the worse trade.

Scenario C — Catching an AI-readiness cliff

Setup. Supplements brand. Agent-ready percentage on the Pulse dropped from 81% to 44% in one day.

Cause. A bulk catalog import had written empty vendor and product-type fields across 1,200 products. Nothing on the storefront looked broken — the products still rendered fine.

What surfaced. 1,200 readiness findings collapsed by dedupe into one per product, with severity escalating to critical on the subset that had also lost price data — those were the ones an AI agent could no longer sell at all.

Fix. Re-ran the import with correct field mapping. Readiness recovered to 83% within two syncs, and the findings auto-resolved on the 7-day window without anyone closing them by hand.

Scenario D — Dismissal as a real signal

Setup. Furniture brand kept receiving replenishment findings for a discontinued product line.

Action. The merchandiser dismissed them with the line marked end-of-life. Dismissed findings leave the worklist and stop counting toward the money rollup.

Why it matters. The "Recoverable now" figure dropped by $4,100 — and that was the correct number. Overstating recoverable money is the fastest way to make the headline worthless, which is why dismissals subtract from it immediately rather than being hidden as "closed but still counted."

Best practices

✅ Work top-down and stop when you're out of time. The list is already sorted by value-per-effort. Working it in order is mathematically the best use of a fixed hour.

✅ Use Weekly Plan for the team, Actions for yourself. Five curated items are a plan people execute; a 34-row queue is a backlog people avoid.

✅ Dismiss aggressively and honestly. A dismissed finding is information — it corrects the money rollup and keeps the headline number trustworthy.

✅ Read the Pulse before the tabs. If agent-ready % or recoverable moved sharply overnight, that's your day, regardless of what's at the top of the list.

✅ Give teammates growth:read:all liberally, growth:manage:all deliberately. Everyone benefits from seeing the queue; status changes should have a named owner.

✅ Bookmark the tab you live in. ?tab=leaks and ?tab=readiness are stable URLs — good for a browser pin or a recurring standup link.

❌ Don't treat "Recoverable now" as forecast revenue. It's a prioritization projection with conservative recovery rates applied, not money you will bank. Realized revenue is reconciled separately through attribution.

❌ Don't chase the largest number on the page. Effort and confidence are in the ranking for a reason; the ordering already accounts for what the raw figure hides.

❌ Don't manually groom old findings. Auto-resolution handles staleness per domain. Time spent closing 12-day-old findings is time spent on rows that were about to close themselves.

❌ Don't ignore an empty list. No open actions is a real signal that your engines are current — not a sign something is broken.

Plan tiers

Capability	Free	Starter	Pro	Agency	Enterprise
Actions worklist	—	✓	✓	✓	✓
Executive Pulse	—	✓	✓	✓	✓
Revenue Leaks tab	—	✓	✓	✓	✓
AI Readiness tab	—	✓	✓	✓	✓
Pages (GSC) tab	—	—	✓	✓	✓
Weekly Plan (Growth Operator)	—	—	✓	✓	✓
One-click fixes from the worklist	—	—	✓	✓	✓
Brand authority in the Pulse	—	—	✓	✓	✓
Role-based view/act separation	—	—	✓	✓	✓
Cross-store roll-up	—	—	—	✓	✓

Frequently asked

How often does the list refresh? Each engine writes on its own schedule — leak detection and readiness scoring run with catalog and order syncs, content decay and voice run on slower cycles. The page reads live on every load, so you always see the current state rather than a cached snapshot.

Why did a finding disappear without me touching it? Either it auto-resolved on its domain window, or the underlying fact changed — the cart converted, stock came back, the price returned inside guard. Both are recorded with a resolution timestamp.

Can I see findings I've already dismissed? Dismissed and actioned findings leave the worklist but are retained with their status, actor, and timestamp. They're excluded from the money rollup because they're no longer recoverable.

Does the worklist show everything, or just a page? The full actionable set is read and ranked, then served a page at a time — 10, 20 (default), or 50 rows. Ranking always happens across the whole set, so page 1 is genuinely the top of the list, not the top of a sample.

What happens if I action the last item on a page? The view falls back to the last valid page rather than showing you a blank screen.

Why are there two money numbers that don't add up? Because they're different things. "Revenue driven" is realized and attributed. "Recoverable now" is a projection of what's still on the table. Adding them would double-count intent as income.

Can the platform just fix these automatically? Findings with a wired one-click fix can be executed from the list. The Weekly Plan deliberately ships in recommend-and-approve mode — it proposes and sequences, you approve. Fully autonomous execution is gated behind Automations and its approval queues.

Do I need to configure anything to start? No. Every engine you have enabled emits into the queue automatically. The list populates as your engines run.

See also

• Actions worklist — the ranked queue in detail
• Weekly Plan — the curated top-5
• Revenue Leaks — recoverable money, honestly rated
• AI Readiness — can agents sell your products?
• Pages performance — search traffic by page
• Command Centre & dashboards — the reporting views this complements
• Decision Intelligence — strategic recommendations layered on top
• Automations — turning repeated actions into rules