SLA + Performance
The numbers we publish (and the ones we don't).
Updated weekly. Aggregated across 26 city nodes. If we miss an SLA we tell you in the next dashboard, not the next sales call.
Updated today
The 6 SLAs we commit to
What we measure. What we commit to. What we hit.
Every number on this dashboard is a rolling 30-day median (or share, where the metric is a percentage), aggregated across our 26 city nodes. Status dots reflect this week's read.
-
Pick & pack accuracy
0 %Commit floor: 99.5%Units shipped correctly as a share of total units picked, audited against the order record at pack.
-
Receipt-to-pickable
0 hCommit ceiling: < 24hMedian hours from dock-receive scan to the SKU being live in the pickable inventory feed.
-
Same-shift cutoff adherence
0 %Commit floor: 98%Share of orders received before the local cutoff that leave the building on the same shift.
-
Exception queue depth
0 %Commit ceiling: < 2%Share of open orders sitting in an exception queue (address fail, inventory mismatch, hazmat hold) longer than 4 hours.
-
Carrier first-scan latency
0Commit ceiling: < 4hMedian time from carrier label printed to first carrier scan, measured across UPS, FedEx, USPS, Purolator, and Canada Post.
-
Inventory accuracy
0 %Commit floor: 99%Cycle-count agreement between system inventory and physical count, rolled across every SKU monthly.
The metrics we refuse to publish
The 4 numbers we keep off the dashboard.
Some metrics look great in a deck but are gameable, brand-mix dependent, or not a Vertex variable. We refuse to publish them. Here is what they are, and why.
- 01
Total order cycle time
Mostly carrier-and-destination dependent. A 4-day median on a Toronto-to-Halifax shipment is not a Vertex variable, it is a Canada Post route. Publishing it would let us take credit for carrier performance we do not own, and would penalize us for routes we cannot speed up.
- 02
Customer satisfaction score (CSAT)
Measured by the brand at the brand-customer interface, not by us. We can influence it (right product, on time, undamaged) but we cannot measure it without seeing the brand's support tickets. If you want to wire your CSAT into our ops review, we will read it weekly. Just not on a dashboard we control.
- 03
Return-rate by SKU
A brand-quality and product-fit metric, not a fulfillment metric. A 14 percent return rate on a swimwear brand tells you nothing about our pick accuracy. We surface returns data inside the brand's ops review, we do not publish it as a Vertex performance number.
- 04
Average cost per order
Too brand-mix dependent to publish across cities. A subscription brand shipping 1-unit boxes at 8 oz is on a different cost curve than a B2B brand shipping 18-unit pallets. We publish line-item pricing for every fee, and we benchmark a brand's cost-per-order against their own prior period, not against an aggregate that means nothing.
City-level performance
26 city nodes. Last 30 days.
Pick accuracy and cutoff adherence at every metro we run. Yellow flags name themselves. If a city is in red, we list it here before we list it in our quarterly board pack.
- Los Angeles 99.98% 99.7%
- Atlanta 99.96% 99.4%
- Toronto 99.97% 99.6%
- Miami 99.93% 98.9%
- Houston 99.94% 98.2%
- Dallas 99.97% 99.5%
- New York 99.95% 99.1%
- Vancouver 99.98% 99.8%
- Denver 99.96% 99.3%
- Nashville 99.97% 99.5%
- Philadelphia 99.94% 99.0%
- Las Vegas 99.96% 99.4%
- Austin 99.97% 99.6%
- Boston 99.95% 99.2%
- Chicago 99.97% 99.5%
- Seattle 99.98% 99.7%
- Phoenix 99.96% 99.4%
- Detroit 98.90% 99.1%
- Calgary 99.97% 99.6%
- San Diego 99.96% 99.5%
- San Francisco 99.95% 99.3%
- Charlotte 99.96% 99.4%
- Portland 99.97% 99.5%
- Oklahoma City 99.96% 99.3%
- San Antonio 99.97% 99.4%
- Montreal 99.97% 99.6%
Incident log · last 90 days
What broke. What we did.
We log every incident that interrupted operations for more than a single shift. Internal-cause incidents go first. External-cause incidents come with the resolution and the policy change we made afterward.
- 2026-04-12 External
Port of Vancouver carrier strike
A 48-hour drayage labor action at the Port of Vancouver delayed 18 Canadian-bound shipments. We refunded affected brands' outbound freight costs ($4,840 across 4 brands), re-routed urgent SKUs through the Calgary node, and held a post-incident review on contingency drayage capacity at Deltaport. New playbook: 5-day pre-book minimum on all Vancouver inbound through Q3.
- 2026-03-03 Internal
Houston WMS deploy window
A scheduled Datex deploy at our Houston node held FBA outbound for 4 hours past the planned maintenance window. Caught on internal monitoring at 10:47 AM CT, no brand-visible miss because the deploy ran inside the same-day cutoff buffer. We have since moved all Houston deploys to Saturday 4 AM CT to give a 12-hour rollback window.
- 2026-02-20 External
Section 122 cross-border surcharge
CBP's Section 122 surcharge implementation created emergency duty-recalc requirements for 6 brands shipping bi-directionally on the US-Canada lane. We absorbed the brokerage re-file costs ($1,260 total), held a 72-hour war room with our broker, and shipped updated landed-cost models to every affected brand inside 48 hours.
Internal monitoring
The metrics we watch every shift but keep off the public dashboard.
We monitor a deeper stack of operational signals internally. They are not on this page because they need context to read correctly, not because we hide them. Any brand under contract can request the full pack at any quarterly review.
- Receiving queue depth by hour
- Dock-seal count vs ASN
- Pick-path velocity per labor hour
- Pack-station throughput
- Carrier manifest cutoff slip
- Claim resolution time by carrier
- Cross-border brokerage cycle time
- FBA inbound appointment hit rate
- Inventory shrink by SKU class
How we calculate
The math behind every number on this page.
Honest dashboards stand or fall on methodology. Here is what we count, how, and what we exclude.
- Rolling window
- Every percentage and median on this dashboard reflects a trailing 30 calendar days, refreshed Monday at 6 AM PT. We do not cherry-pick a peak-week or trough-week window.
- Sample sizes
- Each city node contributes between 4,200 and 38,500 monthly outbound orders to the aggregate. Pick accuracy is sampled at pack and validated via post-ship audit on a 1.5 percent randomized sample.
- Exclusions
- We exclude force-majeure events (carrier strikes, regional weather declarations, port closures) from the cutoff-adherence number, and we name them in the incident log instead. We do not exclude internal-cause misses.
- Audit
- Numbers reconcile to our Datex WMS exports. Any brand under contract can request the raw export tied to their account at any time. The aggregate published here is the same number we read in our internal Monday ops review.
Have an SLA question? Get the full performance pack.
We send the full internal monitoring stack (queue depths, dock-seal counts, claim resolution times, FBA appointment hit rates) to any brand that asks. No NDA gate. No sales script.