Survey 1 · Internal failure analysis
What did we do wrong—
and what could have saved the client?
The customer’s stated reason is evidence. It is not automatically the diagnosis.
We reviewed the full operating trail across Sales, Development, PM, Billing, email, and tickets—then separated the cancellation trigger from the internal failures that may have made the account vulnerable.
01 · The new lens
One cancellation. Four different truths.
The site now makes the management question explicit. We do not stop at the dropdown reason or the final customer sentence.
What the customer said
The direct claim, in their own words when available.What the record proves
Promises, build history, tickets, messages, billing, and outcomes.What failed internally
A supported or possible Sales, Development, PM, or Billing issue.What might have saved them
The earliest realistic intervention—not a hindsight fantasy.02 · Preventability
We found failures. We did not always prove causation.
Six Website deal rows contained at least one supported internal process issue. Only four had enough evidence to connect the situation to a plausible save opportunity.
Preventability across 19 Website deal rows
4114
Partly preventable 4Not evidently preventable 1Unknown 14
6supported internal issue observed
≠4partly preventable cancellation
A billing, QA, or routing defect may exist even when the final cancellation cause cannot be established.
Where the operating record points
Process findings by function
Deal-row classifications across 19 Website cancellations. These are not employee rankings.
Sales
Development
PM / recovery
Billing
Supported issuePossible issueNo issue observedUnknown
03 · Case evidence
What we did, what mattered, and where the save lived.
These are anonymized case-level diagnoses. They illustrate the method; they do not establish prevalence across the company.
Case 11303 days after delivery
We never reconciled activity with value
Customer said
The customer reported little commercial return, unfinished corrections, and a move to a new provider.
Evidence shows
Long and reopened tickets, pricing confusion, competing completion claims, and no raw qualified-lead or revenue truth set.
Our diagnosis
A shared failure: unclear expectations, slow follow-through, and no agreed definition of a successful lead.
Save opportunity
A 30/60-day value review, signed correction closure, and lead-quality reconciliation before frustration hardened into a switch.
Case 1556 days after delivery
The stop signal never became a save signal
Customer said
The customer said a partner preferred other website developers and asked what was needed to place the account on hold.
Evidence shows
A later draft-review follow-up was still sent. No acknowledgment, routing record, or recovery attempt was preserved.
Our diagnosis
The partner decision may not have been ours—but the missed routing removed our chance to understand or recover it.
Save opportunity
Same-day escalation, one accountable owner, and a partner-objection call within one business day.
Case 011,096 days after delivery
Quality, price, and ownership were fragmented
Customer said
Records variously pointed to cost, another provider, an incorrect address, an unwanted photo, chargeback, and nonresponse.
Evidence shows
An outside provider was taking over; billing disruptions and rate changes were real; the final website state and handoff were not reconciled.
Our diagnosis
No single owner assembled the full account truth early enough to distinguish a fixable service problem from a completed provider switch.
Save opportunity
Rapid site audit, address/photo correction, billing reconciliation, and a documented recovery decision before releasing the handoff.
Case 18Pre-delivery fallout
We charged before alignment was proven
Customer said
The customer reported limited funds, disputed the upfront charge, and said the Shopify scope and recurring price differed from the order understood.
Evidence shows
The record contains $144, $244, and $394 recurring figures plus conflicting “update” versus “replicate” descriptions.
Our diagnosis
Sales and billing controls failed to preserve one accepted scope, one price, and a verified readiness decision.
Save opportunity
Signed terms before charge, scope read-back, affordability check, and a phased option when the buyer cannot fund the full setup.
Case 021,068 days after delivery
Not every cancellation is ours to prevent
Customer said
The customer said the business was slowing down, praised the team, and said the website was well liked.
Evidence shows
The direct message aligns with the business-change label; no service complaint was established at cancellation.
Our diagnosis
Customer business change—not an evidenced Sales, Development, PM, or Billing failure.
Save opportunity
A pause or low-cost archive offer was worth trying, but retention was not evidently within our control.
04 · Operating response
Turn each failure into a control and a clock.
The pilot supports concrete process changes now—even before a larger cohort supports prevalence claims.
01Sales
One accepted promise
Record the customer goal, scope, platform, timing, setup price, monthly price, financing, and lead-language read-back before charging.
Before setup payment02Development
Evidence-backed launch
Preserve QA, live-site proof, customer revisions, form delivery, final approval, and any unresolved defect at handoff.
Before Delivered03PM
A recovery clock
Route every stop, hold, dissatisfaction, lead-value, or provider-switch signal to one owner within one business day.
At first risk signal04PM + SEO
Define value together
At 30, 60, and 90 days, reconcile traffic with qualified calls, forms, lead quality, response handling, and the customer’s actual outcome.
After launch05Billing
One commercial truth
Maintain the accepted price, every approved change, successful charges, credits, disputes, refunds, and final account closure in one linked record.
Every price changeProposed save sequence
1Risk signal2Owner assigned3Account truth assembled4Recovery offer5Outcome documented
Target: first owner response within one business day.05 · Evidence discipline
“Unknown” means investigate—not exonerate.
Fourteen of 19 Website deal rows remain unknown on preventability. That is the largest operational finding in Survey 1.
Preserve it exactly, but test it against the full journey.
Separate staff completion claims from customer acceptance and live-state proof.
Rank causes only when the evidence connects the failure to the decision.
Name the earliest realistic intervention and the owner who could have acted.
06 · Scale plan
Ask the same accountability question 100 times.
The next cohort should be the most recent 100 cancellations, reviewed with the same internal-failure, preventability, and save-opportunity framework. Then expand to the last three months—estimated around 900.
Prove the method
Separate claims, triggers, internal failures, causation, and save opportunities.
Find repeatable failures
Test whether the same process breakdowns recur across recent cases.
Measure the patterns
Review the full three-month period after the 100-case quality gate passes.