The situation
One of Common Ground's portfolio companies bids for owner's-representative and construction-management work with public agencies. That work is posted across more than a dozen bid portals, most of them walled behind a login, with no single feed a small firm can watch by hand every morning. We built a daily sweep that reads the open and logged-in portals and scores every hit GO or NO-GO against a defined, licensed positioning, with the reasoning written down hit by hit.
Clearing that first filter is not the end of the problem. A public RFP's required sections look uniform on the page, nine numbered boxes in a fixed order, but they are not uniform underneath. Most of a proposal responds to more writing hours. Some of it is a fact about a relationship that already exists or does not, and no amount of drafting time changes that fact. One real pursuit, a $65 to $75 million public affordable-housing RFP, exposed that gap directly: a full nine-section, rubric-matched proposal was built in two days, and it was never submitted, because a reference requirement worth twenty of the rubric's hundred points was a supply problem, not a writing problem, and it was checked fifth instead of first.
That pursuit produced a reusable proposal method, an itemized fee-translation approach, and a two-question go/no-go gate that now runs before a single section gets drafted on the next one.
Public bid work for owner's-representative and construction-management services is scattered across a dozen-plus government bid portals, most walled behind logins that a curl or a script cannot read; and once a bid clears that filter, a proposal's nine required sections look equally weighted on the page when they are not: some respond to more writing hours, and at least one is a fact about a relationship that either exists or does not.
Built a daily sweep across the open and logged-in portals that scores every hit GO or NO-GO against a defined positioning, with the reasoning written down; built a proposal method that allocates drafting effort by the rubric's own point weight; and, after one real pursuit exposed the gap between a writing task and a supply task, built a two-question gate that checks the fact-based requirement before any section gets drafted.
One sweep sorted nine solicitations GO and six NO-GO in a single pass; a second sweep the same week flagged a bid that had already lapsed its own mandatory pre-bid date, which is now checked at flag time rather than after. A separate nine-section proposal for a public affordable-housing RFP was built end to end in two days and never submitted, on a five-minute check that should have run first. The proposal method and the itemized fee-translation approach moved into the standing toolkit; the two-question gate now runs before that work starts again.
Read on
The decision and the reasoning
Why we did it this way, told first.
ReadWhat we did and what it produced
The work, decision by decision, and the result.
ReadA slice of the project list
A few related projects.
ReadPrivate side
The full report, the source files on record, the timeline and every figure with its source. Password.
Enter