Diagnostic · 12 min
Is your problem traffic, conversion, sales, or operations?
It is easy to arrive with a proposed fix before the underlying constraint has been measured. This diagnostic reverses that sequence: test the path first, then choose the work.
When revenue is flat, more marketing is one possible answer. It is not the only one: demand may already exist while the path loses inquiries or decisions somewhere unmeasured. The tests below distinguish those cases before additional demand is purchased.
This is a diagnostic you can run yourself in an afternoon. Nine checkpoints, each with a symptom, a test you can perform, and what fixing it involves. Work through them in order and use records rather than recollection, especially for routing and follow-up.
Why order matters
The checkpoints are sequential, and a failure early in the chain masks everything after it. If nobody finds you, you cannot tell whether your website converts. If your website does not convert, you cannot tell whether your follow-up works, because there is nothing to follow up.
So work forwards, and stop at the first checkpoint that clearly fails. Treat it as the candidate constraint. Fixing anything downstream of it may not show its full value until the constraint moves.
The nine tests
01 · Demand
“Hardly anyone finds us.”
How it fails. Invisible in search, wrong keywords, wrong audience.
The test. Open Search Console. What share of your clicks are for your own business name? If it is most of them, people are finding you because they already knew about you.
If it fails. Checkpoint 1 work: technical foundations, then content built around the questions buyers actually search before they buy. Organic work needs a multi-month measurement plan rather than a promised ranking date. SEO and AI visibility.
02 · Conversion
“We get traffic. Almost nobody contacts us.”
How it fails. Unclear offer, slow pages, no obvious next step, no trust signals.
The test. Compare relevant landing-page visits with qualified inquiry events. If material, relevant traffic produces few or no qualified actions, conversion is a live suspect. There is no universal percentage threshold.
If it fails. Checkpoint 2 work: positioning, page architecture, and the specific information a buyer needs before contacting a stranger with money attached. Conversion optimization.
03 · Capture
“Leads come in through five different places.”
How it fails. Scattered channels, missed calls, silent form failures.
The test. Count last month’s inquiries from every channel: web form, phone, email, chat, marketplace. If you cannot produce that number in under five minutes, capture is broken.
If it fails. Checkpoint 3 work: unified intake so every inquiry lands in one record with its source attached. Lead generation systems.
04 · Routing
“Good inquiries sit for a day before anyone sees them.”
How it fails. Manual triage, unassigned leads, no response standard.
The test. Take a representative sample of recent inquiries and calculate the median time to first human response from message or CRM timestamps. Compare it with the standard the business intended to meet.
If it fails. Checkpoint 4 work: assignment rules, a response standard, escalation, and a fallback for when the assigned person is away. CRM and sales systems.
05 · Follow-up
“We only chase the ones we remember.”
How it fails. Depends on memory, stops after one attempt.
The test. Review a representative set of lost or stalled deals. How many have a recorded decision, reason, and consent-aware next action? Missing outcomes make silence look like a deliberate loss.
If it fails. Checkpoint 5 work: follow-up sequences that run without anyone remembering, with consent handling built in. CRM and sales systems.
06 · Sales
“Only I can close a deal properly.”
How it fails. The close is undocumented and unrepeatable.
The test. Could someone other than you take a deal from first call to signature this week, using written material that already exists? If not, the process lives in your head.
If it fails. Checkpoint 6 work: documented qualification, discovery, quoting, and handoffs so a salesperson inherits an inspectable process. CRM and sales systems.
07 · Measurement
“I can't tell which channel produced last month's revenue.”
How it fails. No attribution, vanity metrics, disconnected tools.
The test. Name the channel that produced the most revenue last quarter, and the number. If the answer contains the words "I think" or "probably", you do not have attribution.
If it fails. Checkpoint 7 work: source-to-customer attribution at a scale a small business can maintain, and reporting with the definitions written down. Analytics and attribution.
08 · Automation
“The same manual steps consume material time every week.”
How it fails. Manual work, or automation nobody monitors.
The test. List repeated tasks that copy information between systems. Multiply observed frequency by time per occurrence and annualise using the weeks the process actually runs.
If it fails. Checkpoint 8 work: process design first, then automation of suitable steps, with failure alerts and a manual path. Business automation.
09 · Compounding
“We're busy but we've been flat for a year.”
How it fails. No experiments, no baseline, no review.
The test. What did you deliberately change in the last quarter, and what happened as a result? If there is no answer, nothing is compounding.
If it fails. Checkpoint 9 work: baselines, one change at a time, and a review cadence that survives a busy month. Growth experimentation.
What to do with the result
If exactly one checkpoint failed, you may have a focused project. Fix that one thing, measure it, and re-run the diagnostic in a quarter.
If three or more failed, the problem may be the handoffs between them. An individual fix can underperform when the next checkpoint cannot carry the result. That is the case for treating the whole path as one system, which is what growth engineering means.
If none of them clearly failed and revenue is still flat, the constraint is probably not in the growth system at all. It may be pricing, positioning against a specific competitor, or the offer itself. Those are real problems and this diagnostic will not find them.
A costly mismatch to test for
One plausible mismatch is a business with an unmeasured routing or follow-up problem buying more demand. If inquiries already wait without an owner or recorded next action, an SEO engagement can increase the volume entering the same gap without explaining the revenue result.
Response time and follow-up are worth measuring because the data often already exists in message and CRM timestamps. If either is the active constraint, the business can compare a process change against its own baseline before investing in more traffic.