At a glance
Of every Vortex IQ finding written to Crisp in the last 90 days, what percentage has been resolved. The single number that answers “is the team actually fixing the problems Vortex IQ surfaces, or is the audit just generating noise?”. Above 75% is healthy; under 50% means the audit pipeline is creating more work than the team can absorb.
| The formula | resolved_in_90d ÷ created_in_90d × 100, where resolved_in_90d counts conversations that moved to state = resolved between (NOW - 90 days) and NOW, regardless of when they were created; and created_in_90d counts conversations created between (NOW - 90 days) and NOW with the vortexiq:finding segment. |
| Why this shape, not “closed / open ratio” | A “closed vs open” snapshot ratio is heavily distorted by what is currently in the queue. Rate-over-rolling-window normalises against creation cadence, so a brand that creates 50 findings/month and resolves 45 reads identically to one creating 5 and resolving 4.5. Both are healthy operations. |
| Resolution definition | state = resolved only. We do not count archived or deleted conversations as resolved (those typically mean “given up”, not “fixed”). This is stricter than Crisp’s own UI, which conflates the three states under “closed”. |
| Numerator/denominator alignment | Both windows are 90 days from NOW. A finding created 95 days ago and resolved yesterday counts in resolved_in_90d but NOT in created_in_90d. This means the rate can briefly exceed 100% during cleanup pushes (lots of old findings getting fixed). The card clamps display to 100% but the underlying signal is informative. |
| Free-tier note | Crisp Free plan deletes conversations after 30 days, breaking the 90-day denominator. Vortex IQ’s audit ledger backstops the count, so the rate stays accurate, but Free-tier merchants cannot independently verify it. Pro tier recommended for any merchant relying on this card for governance. |
| Why 75% / 50% | Calibrated against ~12 months of merchant data: brands that maintain >75% resolution rate consistently improve revenue and SLA quarter-on-quarter. Brands below 50% accumulate technical debt that compounds; the audit pipeline starts feeling like a chore and adoption drops. 50-75% is the working amber band where most merchants live. |
| Time window | 90D, rolling. |
| Alert trigger | <50%. |
| Sentiment | Gauge with good=75, warn=50. |
| Roles | owner, operations |
Calculation
Calculated automatically from your Crisp data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
A French fashion brand on Shopify Plus uses Crisp Pro across two Crisp websites (one French DTC, one English-EU wholesale). Snapshot taken on 28 Apr 26 at 17:00 CET. The window is 28 Jan 26 to 28 Apr 26. In the last 90 days:| Activity | Count | Where it sits today |
|---|---|---|
| Findings created | 64 | 41 resolved, 18 still open, 5 archived without resolution |
| Findings resolved (created in window) | 41 | Closed within the same 90 days |
| Findings resolved (created earlier, closed in window) | 7 | Old work cleared during a Q1 cleanup push |
| Activity | Count |
|---|---|
| Findings created | 38 |
| Findings resolved | 11 |
| Findings still open at age >30d | 19 |
Sibling cards merchants should reference together
Resolution Rate is the headline efficiency metric. Pair it with the cards that explain why the rate is what it is and whether the resolution is real:| Card | Why pair it with Resolution Rate | What the combination tells you |
|---|---|---|
| VortexIQ Findings Open | The numerator’s complement. | High rate (>80%) but Open >20 = inflow exceeds capacity; rate is healthy but absolute backlog is growing. |
| VortexIQ Findings Resolved | The numerator (raw count). | Rate of 75% is meaningless without volume context. 75% of 4 findings is 3 closes; 75% of 80 is 60 closes. |
| Abandoned Findings (>14d no movement) | The “given up” subset that drags resolution rate down. | High abandoned + low rate = the abandoned ones are killing the rate. Push those first. |
| Avg Time-to-Fix (days) | Speed of closure. | Rate 80% + Time-to-Fix 18d = team closes most things, but slowly; opportunities decay. Rate 60% + Time-to-Fix 4d = team closes fast but ducks the hard ones. |
| Throughput Trend | Rate over time, raw counts. | Confirms whether resolution rate is improving, plateauing, or regressing. |
| Sprint Velocity | Engineering capacity. | Rate < 50% with velocity stable = engineering is busy on non-finding work; not a capacity issue, a prioritisation issue. |
| Shopify / BigCommerce / Adobe Commerce Total Revenue | The dollarised peer. | Brands that maintain >75% rate consistently grow revenue 2-4 percentage points faster quarter-on-quarter than brands below 50%. The card has measurable financial signal. |
| Crisp Open Tickets | Total inbox volume. | Rate 80% + Open Tickets 200 = team is great at audit follow-up but the customer-chat side is drowning. Different problem, same team. |
Reconciling against the vendor’s own dashboard
Where to look in Crisp’s own dashboard: Crisp does not natively report a resolution-rate metric. The closest equivalent requires two manual counts:Crisp App → Inbox → Filters → SegmentFor a programmatic check, Crisp’s Statistics API exposes per-website conversation counts but does not split by tag, so the merchant’s own automation will need to query List Conversations withvortexiq:finding, Date created (last 90 days), State = Resolved Note the count. Then change State toAllto see the total created in the same window. Divide.
search_query=segment:vortexiq:finding and group by state.
Why our number may legitimately differ from a manual Crisp count:
| Reason | Direction | Why |
|---|---|---|
| Resolution definition | Ours stricter (lower) | Crisp UI groups archived, deleted, and resolved under “closed”. We count resolved only. A merchant who archives instead of resolving will see Crisp report higher closure than our card. |
| Window boundary | Negligible | Both compute on a rolling 90-day window. Boundary precision is to the minute. |
| Free-tier history cap | Ours higher | Free Crisp deletes bodies after 30 days, but the audit ledger keeps the IDs. Crisp’s UI count for “created in last 90 days” understates by the deleted bodies; ours does not. |
| Numerator/denominator drift | Either, briefly | Findings created 95 days ago and resolved today count in our numerator but not denominator (the denominator window starts at NOW - 90 days). During cleanup pushes the rate can briefly read >100%; the card display clamps to 100%. |
| Time zone | Negligible | UTC versus workspace timezone shifts the boundary by at most 12 hours; over a 90-day window this changes the rate by <1 percentage point. |
| Card | Expected relationship | What causes the divergence |
|---|---|---|
liveagent.liv_vortexiq_finding_resolution_rate / livechat.liv_vortexiq_finding_resolution_rate / tidio.tid_vortexiq_finding_resolution_rate | Definitional twins on other live-chat tools. | A merchant running findings on multiple chat tools should expect similar rates if the team is the same; very different rates by tool means routing rules favour one tool over the other. |
zendesk.zen_vortexiq_finding_resolution_rate / freshdesk.fd_vortexiq_finding_resolution_rate | Helpdesk peers. | Async helpdesk rate often runs 5-10 percentage points lower than chat rate, because async tools have softer SLA expectations. |
shopify.total_revenue / bigcommerce.total_revenue / adobe_commerce.total_revenue | Statistical correlation: maintaining >75% rate over 4 quarters has shown 2-4pp faster revenue growth than peers. | The mechanism is straightforward, audit findings name real merchant problems; closing them removes friction. |
datadog.dd_health_score | Operational-health peer. | Different layer (Datadog measures live infra, this card measures team follow-through), but the brands that score well on both are usually the same brands. |
Known limitations / merchant FAQs
My rate is 90% but my Open count is 35. Is that good or bad? Depends on your inflow rate. If 35 open while resolving 9 of every 10 you create, the absolute pile is growing because you are creating more than 9 a week. The rate metric is healthy; the Open Findings absolute count is the lever to pull. Either lower inflow (turn off low-priority audit modules in Vortex IQ → Settings → Audit) or add capacity. Rate alone is not enough; track both. My rate dropped from 80% to 60% in a single week. What happened? Three usual causes, in likelihood order. (1) Bulk-created findings hit the inbox after a Vortex IQ audit module was newly enabled (e.g., the team turned on the Klaviyo audit, which created 12 new findings overnight). The denominator jumped, the rate dipped temporarily; this self-corrects within 30-60 days. (2) Engineering capacity diverted to a feature launch; throughput halved. (3) A specific finding type the team cannot resolve (e.g., third-party platform issue) is accumulating and pulling the rate down. Open Throughput Trend to see which. How is “resolved” different from Crisp’s “closed” or “archived”? Crisp has three terminal states:resolved, archived, and deleted. Resolved means the agent actively clicked Resolve, signalling the work is done. Archived typically means “we are not going to action this” (often used for spam or off-topic). Deleted is housekeeping. We count only resolved because the question this card answers is “is the team fixing problems”, not “is the team clearing the inbox”. Bulk-archiving will NOT improve this rate, which is intentional.
Can I exclude specific finding types from this rate?
Yes, in two ways. (1) Module level: turn off the audit module that creates them (Vortex IQ → Settings → Audit → toggle off). (2) Tag level: rename the segment from vortexiq:finding to something else (e.g., vortexiq:deferred); they leave the count entirely. Use option (2) if you want to keep the conversations visible in Crisp but drop them from rate maths.
Why is the rate sometimes >100% and clamped to 100%?
During cleanup pushes when the team resolves a batch of old findings (created >90 days ago, so outside the denominator window), the numerator includes them but the denominator does not. Rate temporarily reads e.g. 110%. We clamp display to 100% but it is a useful informal signal that a cleanup is happening. The number self-corrects as the cleanup completes.
My team uses Crisp Free plan. Does that affect this card’s accuracy?
The rate is accurate because Vortex IQ’s audit ledger backstops both numerator and denominator. What you lose on Free is the ability to inspect the underlying conversations in Crisp directly past 30 days; the count itself is intact. Pro tier ($25/mo) is recommended if this card matters for your governance.
Should the rate be 100% all the time?
No. A consistent 100% means the team is resolving everything, which usually means findings are being closed mechanically (someone hits Resolve without fixing) rather than thoughtfully. The healthy band is 70-85%; that range reflects normal triage where 15-30% of findings are legitimately deferred (won’t fix, depends on platform vendor, deferred to next quarter). Below 50% is the problem zone; above 95% may be a hygiene signal worth investigating.
Can I see this rate per agent?
Not on the card itself, but Crisp’s Statistics → Performance view shows resolved-conversation count per agent. Cross-reference with Tickets by Assignee to see whose rate is dragging the team rate down. That conversation should happen in private; rate is a leading indicator of burnout, not a performance review tool.
My team is using Crisp’s chatbot to auto-resolve findings. Is that gaming the rate?
Yes, if the bot is closing findings without a human ever engaging. The rate would read healthy while the merchant problems remain. Two safeguards: (1) we exclude bot-only resolutions from the numerator if the conversation has zero human messages (we detect this via Crisp’s List Messages endpoint); (2) pair this card with Avg Time-to-Fix, bot-resolutions show abnormally low time-to-fix (seconds, not days), which is a tell.
The card refreshes every 60 seconds. Do small changes show up immediately?
Yes, but the 90-day denominator means a single resolution moves the rate by less than 1 percentage point on most merchant accounts. Bulk closures (5+ in one session) are visible within 60 seconds; one-off closures register but do not visibly move the gauge.
Is there a “rolling 30 days” view of this rate?
Not on this card. The 90-day window was chosen because shorter windows are too noisy on smaller merchants (a brand creating 5 findings/month would have a 30-day rate that swings between 0% and 200% based on a single closure). If a merchant operates at high finding-volume (>10/week), the 30-day variant is on the roadmap.