Owner-picked baseline + 90-day criterion
When. You can freeze a number on day 0
You give. The vendor may refuse a fair test
Hotels only · No pay-to-rank · Advice first; verified when attested
Job brief
How and who for the hotel job—verified when attested.
Usable on this screen
Pick one contract. Rebuild the promise as a number you can observe in 90 days. Pull the baseline from the PMS or the utility, not from the vendor QBR. Ashcroft sits on the owner side. A kill is a successful job.
Step 1
If they sold ‘efficiency,’ write the labor hours or the kWh. If they sold ‘direct mix,’ write contribution after commission.
Step 2
Same season, weather-normalized if energy, occupancy-held if mix. The vendor does not pick the baseline.
Step 3
What would have happened anyway. What you cannot attribute. Harbor will not list a proof without a caveat.
Step 4
Renew, cut, or renegotiate. Shopping operators in the same breath is only leverage if you mean it.
Provider shortlist
Owner asset management
Catalog score 8.8. A post-job shape, not a verification. Stable id ashcroft.
Hotel accounting & night audit export
Catalog score 7.8. A post-job shape, not a verification. Stable id ledgerlight.
Boutique acquisition underwriting
Catalog score 8.9. A post-job shape, not a verification. Stable id sable-desk.
Leave a record
Harbor by Signalry stays hotels only. A description, a claim, and a reported result each leave a stable id behind. Nothing here is a price, a customer, or a verification.
Buyer
No account. Harbor by Signalry opens the closest catalog brief, when not to buy, and a shortlist. A weak sentence stays on the catalog.
Open a briefSupplier
Company, jobs served, regions, and a hotel reference. The packet is queued as unverified pending. The form cannot mark it verified.
Start a claimHotel
A metric, before and after, and the dates. It is shown as reported until a person checks the evidence. Verified Outcome is not insurance.
Record a resultAdviceNo baseline, no ROI. A percentage without a window is a brochure.
ClaimedMost ‘3x ROI’ slides skip the counterfactual.
Situation
A vendor QBR with a baseline the vendor picked. The IC wants proof it paid for itself. The house cannot name the start number.
Approaches
When. You can freeze a number on day 0
You give. The vendor may refuse a fair test
When. You want a slide
You give. You will re-buy the same story
When. The contract is not signed yet
You give. A delayed start date
Who to involve
First 7 days
YAML-shaped: day / theme / tasks / owner / risk / exit. Advice until a case is verified.
day: 1
day: 2
day: 3
day: 4
day: 5
day: 6
day: 7
Decision worksheet
Take this to a vendor call. Budget bands are Advice — a Harbor shape, not a quote.
Ask the vendor
Red flags
When not to buy. Do not hire the vendor to grade their own homework. A referral-fee advisor is not an advisor.
Do it in-house
Time and a named owner. No vendor. Use the 7-day plan first.
$8–25k review
Typical Harbor band for this job — advice, not a quote.
Full program
People + system + 90-day proof. Only if the 7-day plan already failed.
Vendor checklist
Weighted questions — yes/no, scale, or text. Must-pass ids persist on this browser. Pass threshold 70. A no on a must-pass is a walk-away — not a maybe.
property context · weight 2 · must pass
Can they name your key count and whether you are independent, soft-brand, or major — without a discovery theater?
property context · weight 2 · must pass
Do they treat your PMS as a shape (Opera-class / Mews-class / Cloudbeds-class) — not a fake partnership slide?
control · weight 3 · must pass
Who types the status / number — a person, or your product?
vendor · weight 2
What would have happened anyway in 90 days without you?
vendor · weight 2
Show a property of this size and stack. Demo-labeled categories do not count as proof.
vendor · weight 2
What can your agent never do (folio post, mark clean, change a rate)?
vendor · weight 2
What is the kill-switch if guest comments move?
Does not clear
0 / 100
must_pass: prove-vendor-roi-90d-ctx-rooms, prove-vendor-roi-90d-ctx-stack, prove-vendor-roi-90d-q1
A must-pass failed. Walk away.
How
Pick one contract. Rebuild the promise as a number you can observe in 90 days. Pull the baseline from the PMS or the utility, not from the vendor QBR. Ashcroft sits on the owner side. A kill is a successful job.
Step 1
If they sold ‘efficiency,’ write the labor hours or the kWh. If they sold ‘direct mix,’ write contribution after commission.
Step 2
Same season, weather-normalized if energy, occupancy-held if mix. The vendor does not pick the baseline.
Step 3
What would have happened anyway. What you cannot attribute. Harbor will not list a proof without a caveat.
Step 4
Renew, cut, or renegotiate. Shopping operators in the same breath is only leverage if you mean it.
Owner-side 90-day memo
You get. A number you can take to IC
You give. Cash and a vendor who may not like the answer
Vendor QBR
You get. Lunch
You give. The baseline
Pitfalls
Success criteria
Who
Rank is post-job score, then listed order. Not for sale. Software marked demo is an archetype.
Independent owner advisors for GOP, management agreements, and capex sequencing — not a brand desk.
$25–60k engagement or $8k/month retainer
Post-job
22 reviews
Buy-side hotel underwriting for 40–200 key independents — ops reality, not just a trailing twelve.
$35–75k buy-side memo
Post-job
11 reviews
Demo-labeled hotel ledger that sits cleanly on a PMS export — not a generic SMB book tool.
Demo · $250–600 / month
Post-job
9 reviews
None ranked in AI agents yet. Harbor will not invent a provider.
Loading the compare table…
Compare
People do the work beside the desk. Software is a system — demo-labeled when it is a category shape. Agents draft; they do not post money or mark a room clean.
| Provider | Kind | Score | Fee | Pick when | Walk away when |
|---|---|---|---|---|---|
| Ashcroft Asset Advisors | People | 8.8 | $25–60k engagement or $8k/month retainer | Owners who want a second set of books that is not the manager's. | Branded flags looking for a PIP vendor. |
| Sable Underwriting Desk | People | 8.9 | $35–75k buy-side memo | First-time or small-fund buyers of a single boutique. | Institutional portfolios with an in-house desk. |
| Ledgerlight Accounting | Software· demo | 7.8 | Demo · $250–600 / month | Owner-operators tired of a generic bookkeeper file. | Multi-brand portfolios with a shared-service finance team. |
Open graph
Evidence
The Annex · 164 keys · Chicago, IL
Read the case
VerifiedGlasswater · 310 keys · Miami, FL
Read the case
VerifiedCompare up to three
Default sort is fit, high to low. Unpublished SLA / term / switch cost is labeled ask — not invented. Demo software is a shape.
Sort
3 of 3 compared. Highest fit is Sable Underwriting Desk at 8.8.
Highest fit
8.8