Team and organization
Team and organization work, where the problem is the operating process rather than any individual analyst.
Diagnosis
Organizations rarely arrive saying they have a decision problem. They arrive with a throughput problem. Analyses get produced, circulated, and admired, and the replenishment rule, the allocation logic, and the expedite heuristic all continue exactly as before. Somebody eventually asks what the analytics function is for, and nobody has a good answer, because the honest answer requires a count nobody has run.
The instinct is to fix the output: better visualizations, tighter packs, a decision log, a template. These produce modest gains, because the output is not where the failure originates. A study changes nothing when the brief it answered was a topic with a deadline attached, and no amount of craft applied at the end recovers a choice that was never named at the beginning.
The tractable intervention is upstream and it is cheap. Change what a request has to contain before work starts, and change what a recommendation has to specify before it counts as finished. Those two changes move more than a year of tooling investment, and neither requires anyone to become a better analyst.
Engagements
Find out what your analytics function is actually producing.
From $15,000
Three to four weeks
Most requested
Teach the method against live work.
From $12,000
Two or three half-days
Get one rule into production, properly.
From $5,000 per month
Monthly retainer, three-month minimum
Pricing depends on organization size, the number of studies in scope, and whether the engagement includes building an internal standard you will maintain afterwards. All figures are starting points for a scoping conversation.
Method
An audit is not a survey of opinion about analytics quality. It is a structured read of real briefs, studies, and write-ups against explicit tests, which is what makes the findings arguable rather than impressionistic.
A function with a mean score of 60 percent might have every study at 60, or half at 95 and half at 25. These are entirely different problems with entirely different remedies. The first is a standards problem and responds to training. The second is a routing problem and responds to changing which requests get accepted, which is cheaper and nobody suggests it.
One page: the choice, the alternatives, the recommendation, the reasons that could change the ranking, the flip conditions, and the rule. An engagement about this method that delivered a forty-seven-slide deck would have refuted itself. The evidence sits behind the page for the people who want it.
The deliverable includes the write-up standard, review criteria, and a short internal guide in your own vocabulary, referencing your systems and your examples. The measure of success is that the standard is still applied two years later, by people who never met me.
Next step
Send three or four representative studies or briefs, redacted as needed. I will come back with an initial read on which of the three parts your work is missing, at no charge and with no obligation.