How we work
A diagnosis names one constraint. A list of forty findings does not.
Audit tools are not diagnostic tools. They are detection tools. The gap between detection and diagnosis is where most SEO reporting fails — and where client trust is won or lost.
The standard approach
What a crawler export actually tells you
A crawler export is useful. It tells you which issues exist on a site and, if ranked by count or estimated impact, which issues occur most frequently. That’s not the same as telling you which issue is causing the metric shortfall you’re trying to close.
The confusion runs like this: a site has 23 pages with missing meta descriptions. That’s a finding. But if the site’s organic traffic deficit is caused by keyword targeting misalignment — competing for head terms it can’t rank for — then fixing all 23 meta descriptions moves the metric by nearly zero. The list has no mechanism. It can’t tell you which issue, if fixed, would change the outcome.
Sending a client a list of forty findings is sending them a problem. A diagnosis sends them a decision — one thing to do next, with an expectation of what will change if they do it.
Clients pay agencies to make hard problems smaller. A forty-item audit doesn’t do that. It redistributes the hard problem from the agency to the client.
The constraint concept
What makes something a constraint
A constraint is the single factor that, if changed, produces the largest improvement in the target metric. It is not necessarily the most common problem, the most severe technical violation, or the first thing on the list. It is the binding factor — the thing the system is limited by right now.
In any website’s organic performance, there are contributing factors to every metric: impressions, click-through rate, conversions, indexed pages, authority, keyword targeting, content quality, technical accessibility. Each of these is a driver. One driver is underperforming by more than the others, relative to what’s achievable. That driver is the constraint.
The goal of a diagnostic is to identify which driver it is, quantify the gap, explain why it exists, and prescribe the minimum intervention that addresses it. Not the complete list of all interventions that would be beneficial. The minimum effective set.
Three fixes is not a round number. It reflects a real constraint on client attention and execution capacity. Twelve recommendations produce twelve partially-executed tasks. Three produce three completed ones.
The method in practice
Driver decomposition — finding the constraint by elimination
Every metric is a product of its component drivers. Decompose the metric, measure each driver against benchmark, and the constraint reveals itself as the one driver performing furthest below its potential.
Example: Organic traffic deficit of −4,200 visits/month
Each driver measured as % of achievable potential given site characteristics. The constraint is the driver with the largest gap.
Impression drivers
Click-through drivers
Conversion drivers
Keyword targeting is performing at 34% of achievable potential — 22–57 points below every other driver. This is the constraint. Fixing index coverage from 91% to 100% recovers 9 percentage points. Fixing keyword targeting from 34% to 60% recovers 26 — and that 60% target is achievable within the site’s current authority ceiling.
Driver percentages are calculated against the achievable range for a site with these specific characteristics — authority, niche, geography, competition density — not against an abstract maximum. A DR 28 niche site at 85% of achievable potential is outperforming a DR 60 site at 34% of its potential.
Output format
Why three fixes, and why sequenced
Once the constraint is identified, the report prescribes three fixes. They are sequenced: the first creates the conditions for the second, which creates the conditions for the third. This is a plan, not a checklist.
The sequencing matters because implementation capacity is finite. The right order means a partial execution still moves the metric. If the client completes fix 1 and stops, they should be meaningfully better off than before. If they complete all three, they’ve closed the constraint.
What each fix includes
- → Specific action, not category
- → Timeframe estimate
- → Expected mechanism — why it works
- → Dependency on prior fix where relevant
What we don’t include
- — Guaranteed ranking positions
- — Traffic or revenue projections
- — Findings we can’t support from outside access
- — Padding to make the report look comprehensive
Stated limits
What external analysis cannot see
Every report includes a limits section. This is not a disclaimer — it’s a diagnostic boundary. External analysis can assess a site’s technical structure, content, indexation, authority, link profile, and competitive positioning. It cannot access server logs, conversion data, CRM data, or business-context information the client holds.
The limits section tells the agency what questions the report can’t answer, so they know what internal data to gather to complete the picture. A diagnosis that overstates its certainty is a liability.
See the method in a real report
The sample report is a complete Constraint Brief on a real site. Ungated.