Evidence-based SWOT
How to use it
- Open Market Intelligence → SWOT and select Build SWOT. It takes about a minute and uses data already in your store.
- Read each quadrant. Every item shows what it's compared against, its impact and confidence, and an Evidence button.
- Triage. Accept items you agree with. Dismiss the ones you don't, with a reason.
- Propose actions on the items worth acting on. A teammate approves or rejects them.
- Run a market report and X-Ray your main competitors, then select Refresh SWOT. Opportunities and threats widen with that evidence.
⏱ ~1 minute to build · 💳 Starter+ · 🎯 A SWOT you can defend in a meeting
Why "evidence-based" matters
A typical AI SWOT reads well and says nothing: "Strong brand identity." "Growing competition." "Opportunity to expand online." Every store in the category gets the same list, and none of it can be checked.
This SWOT applies two rules in code, after the model has drafted its items:
- No item without evidence. Each item must cite at least one verified claim.
- No strength or weakness without a comparison. A 4.6-star rating isn't a strength on its own. A rating that rose from 4.3 to 4.6 over 90 days is. So is pricing below matched competitors. The comparison is what makes it a finding.
Items that fail either rule are dropped before you see the SWOT.
Where the evidence comes from
Your store's measured signals
Strengths and weaknesses can only rest on these. Each has a built-in comparison:
| Signal | Compared against |
|---|---|
| Review rating (90 days) | Your own history |
| Customer satisfaction (90 days) | Your own history |
| Support tickets (30 days) | Your own history |
| View-to-purchase rate (30 days) | Your own history |
| Organic search clicks (28 days) | Your own history |
| SEO health score | Your own history |
| Customer journey score | Your own history |
| AI answer visibility | Your own history |
| Price position | Matched competitors |
| Search share of voice | Competitors |
| Competitor price cuts (30 days) | Competitors |
| Open revenue leaks | Money at risk |
| Stock-out risk | Stock at risk |
A signal is used only when there's enough data for it to mean something. When there isn't, it's listed under Signals not used this time with the reason, for example that there were too few reviews in 90 days to show a trend. A thin number is never used just to fill a quadrant.
"Money at risk" and "stock at risk" need no comparison: a loss is a weakness whatever it's measured against.
Evidence from outside your store
Opportunities and threats must rest on external evidence or on a competitor-comparison signal. When you rebuild the SWOT, it copies in the strongest verified, non-stale findings from:
- your current market reports (the latest two), and
- your current Competitor X-Rays (the latest five).
The findings are copied into the SWOT rather than linked, so each SWOT version stays complete even after the reports it used are replaced.
If you have no market report yet, the SWOT says so: "No market report yet: opportunities and threats rest on competitor evidence only."
Reading an item
| Part | Meaning |
|---|---|
| Title and rationale | What the item is and why |
| Comparison badge | vs your own history, vs matched competitors, money / stock at risk, or external evidence |
| Impact | Low, medium or high |
| Confidence | Never higher than the weakest evidence the item cites |
| Evidence (N) | The claims behind the item, each with its label, support and source |
| Status | Open, accepted, dismissed or actioned |
Two items resting on exactly the same evidence in the same quadrant are merged into one.
TOWS strategies
Below the quadrants, TOWS strategies pair what's inside your store with what's outside it:
| Type | Pairs | Asks |
|---|---|---|
| SO | Strength + Opportunity | How do we use a strength to take an opportunity? |
| WO | Weakness + Opportunity | Which opportunity is worth fixing a weakness for? |
| ST | Strength + Threat | How does a strength blunt a threat? |
| WT | Weakness + Threat | Where are we exposed, and what limits the damage? |
Each strategy has a first step and lists the items it's built on. A strategy is kept only when it really pairs the right quadrants from items that passed. An "SO" strategy built on two strengths is dropped.
Accepting, dismissing and reopening
| Action | What happens |
|---|---|
| Accept | Marks the item as agreed |
| Dismiss… | Asks why. The reason is required and shown on the item |
| Reopen | Returns a dismissed item to open |
Your decisions carry over. When you rebuild the SWOT, an item resting on the same evidence as an item you've already decided on keeps that decision, whether accepted, dismissed or actioned. A dismissed item comes back only when its evidence changes materially. Small wobbles in a figure don't count; the figure is compared at two significant digits.
This is what makes regular rebuilds useful. You triage once, and later SWOTs show you what's actually new.
Proposing an action
Select Propose action on an item. Market Intelligence never changes your store. It can only propose, and a proposal is judged by the same action policy that governs every AI-suggested change in the app.
- Low-confidence items stay recommendations. No approval request is raised, because asking someone to approve a weak guess wastes their attention. You'll see a message saying the evidence isn't strong enough.
- Medium and high-confidence items raise an approval request. The item becomes Accepted with an Awaiting approval badge. The request carries the item, its rationale and its evidence, and is due within 7 days.
- You can't approve your own proposal. Proposals appear under Proposed actions awaiting approval for teammates who can manage the SWOT. If you raised some yourself, the queue says how many are waiting for a teammate.
- Approving records the decision. It marks the item Actioned. It doesn't change prices, products or campaigns. Your team carries out the action.
Plan tiers
| Free | Starter | Pro | Agency | Enterprise | |
|---|---|---|---|---|---|
| SWOT runs per month | — | 4 | 8 | 30 | 100 |
| Opportunities & threats from market reports | — | — | ✓ | ✓ | ✓ |
| Opportunities & threats from X-Rays | — | — | ✓ | ✓ | ✓ |
| Monthly scheduled rebuild | — | — | — | ✓ | ✓ |
On Starter, opportunities and threats come from your competitor price and share-of-voice signals. Market reports and X-Rays, which widen them, start at Pro.
When a SWOT can't be built
If your store doesn't have enough data for any signal yet, for example a new store with few orders, reviews or search data, the run fails with "There is not enough store data yet for an evidence-based SWOT" and lists what's missing. It doesn't produce a generic SWOT instead. Connect Search Console, collect reviews, or add competitors to the price monitor, then try again.
Best practices
- Rebuild after something changes: a price change, a review campaign, a new market report. The carried-over decisions mean you only review what's new.
- Always give a real dismissal reason. It's the record of why the team decided not to act, and it shows next to the item in every later SWOT.
- Propose actions on the items your TOWS strategies are built on first. A strategy that pairs a strength with an opportunity is usually a better use of effort than fixing a low-impact weakness.
- Check the evidence before approving. The proposal carries it for a reason.
- Watch "Signals not used this time". It's a list of what your store isn't measuring well enough yet.
See also
- Market Intelligence overview: labels, budget, plans and permissions
- Market reports: the external evidence behind opportunities and threats
- Competitor X-Ray: competitor evidence for opportunities and threats
- Approval queues: how approvals work across the app
- Revenue Leaks: the source of the money-at-risk signal