Procurement consulting for state, local & education
On $50 million of buying, that is $2 million to $7.5 million a year. GAO found leading agencies recover 4 to 15 percent of services spend. The agency already pays it. It just does not name it.
GAO-13-417, applied to spend you enter
On $50 million of buying, the range is $2 million to $7.5 million a year. That is 4 to 15 percent of services spend — what GAO found leading agencies recover — not 15 percent of the whole appropriation.
Low end · 4%
$2M
High end · 15%
$7.5M
4% of $50 million is $2 million. 15% is $7.5 million. Illustration, not an audit finding.
Seven commercial companies, services they sourced, savings vs. prior year. Not SLED, not a census, not 15% of total budget.
Lower cost to run procurement vs. typical peers. Their $6 million example is functional savings at a $10 billion-revenue company.
Of targeted savings, not of spend. Skipping the path can erase negotiated value. It does not take 15% of the appropriation by itself.
Munivera does not claim your agency is losing $7.5 million. The calculator applies GAO’s published range to spend you enter. Sources: U.S. GAO-13-417; The Hackett Group; published maverick-spend summaries.
The pattern
This is not a maturity model. It is how leakage happens in public agencies: incomplete intake, inbox status, undocumented exceptions, and files that fail audit.
Check every statement that is true in your agency. The cost is already in the budget. It just is not labeled.
Where the money actually goes
Rework, status hunting, and Board-package assembly from email are paid for in fully burdened hours. That cost is real, and it is separate from any strategic-sourcing range. The agency already funds it. It just does not appear as a line item called “broken intake.”
A request that sits incomplete does not only cost the buyer. It costs the department waiting on a contractor, a bus, a roof, or a grant drawdown.
Undocumented exceptions, untraceable files, and informal closeout show up later as protest exposure, audit findings, and a Legal review of the same language for the tenth time.
We do not treat $7.5 million as your number. The 4–15% figure is GAO’s published range on services spend. The diagnostic uses your volume and your procedure. A principal sends a written proposal sized to the work — often to common municipal small-purchase thresholds for the first engagement.
Staff-time sketch
Separate from any 4–15% sourcing range: the loaded hours spent chasing a complete request. This is not a bid and not a savings guarantee. It uses volume, chase minutes, and a loaded rate so leadership can see the current habit before anyone sells software.
Hours / year on chase
—
Loaded cost of that time
—
Do not treat this as a recovery figure. It is the cost of the current habit. A diagnostic uses your real volume and your real procedure.
What replaces it
Bad procurement is usually undocumented procurement. The fix is a complete request, a written path, an identifier from day one, status in the system, and closeout as a required step. Software comes after that is true.
01
A complete requisition, or it does not enter.
02
Thresholds and written procedure choose IFB, RFP, cooperative, sole source, or emergency.
03
An identifier from day one. Status in the system, not an inbox.
04
A named employee signs. The workflow bends to existing authority.
05
The file is finished on purpose, so audit and grantors are not a reconstruction project.
Relevant work
Procurement modernization
RIAC moved from transactional, email-based procurement to a documented, end-to-end workflow: a 15-step process, playbooks for nine paths, PIID numbering, and an 87-page operating guide.
That is the operating layer. Automating an undocumented process only makes confusion move faster.
Also from Munivera
AI operations for the same SLED agencies: a measured pilot in your tenant after the procedure exists. The employee still owns the decision.
Bring the real process
Tell us how requests arrive, who awards, and how you know an action is done. The diagnostic turns that into a measured recommendation — or a clear reason not to proceed.