Case Study 1 — How Kestrel Won Northgate: An Interview, a Scoring Matrix, and Two Promises

All people, companies, and projects in this book are illustrative composites.

Setup

Fourteen months before the notice to proceed you already know — March 3, Year 1 — the Northgate Outpatient Pavilion did not exist as a contract. It existed as a problem on Meridian Health System's capital committee agenda.

Meridian is a regional non-profit health system. Its outpatient clinics had outgrown a leased building on the other side of Rivermont, and the lease on that interim space expired October 1, Year 2. If the new pavilion was not open before then, Meridian would either extend a lease it did not want at a rate it would not like, or move clinical operations twice.

Meridian's board had approved a $61,000,000 total project budget — construction, design fees, medical equipment, furniture, permits, owner's contingency, and financing carry. Priyanka "Pri" Sethi was the owner's representative. Halvorsen + Pike were already engaged as architect. The delivery method was decided early and deliberately: CM at Risk with a guaranteed maximum price, because Meridian wanted a builder in the room during design and a number it could take to its board before construction documents were finished.

That decision set the shape of the competition. Nobody was going to win this job with a price, because there was no price to bid.

What happened

The RFQ

Kestrel's statement of qualifications was competent and unremarkable: healthcare project sheets, an experience modification rate and incident history, audited financials, a surety letter confirming the $150M aggregate / $60M single-project program, and six references. Nine firms submitted something similar. Four went forward.

The shortlist told Kestrel exactly what it was up against. One competitor was a national contractor with a dedicated healthcare group four times the size of Kestrel's whole company. Two were regional firms of roughly Kestrel's size. Nadia Haddad ran the go/no-go on the interview stage — not on whether to submit, which was easy, but on how much to spend. She authorized four people, two Saturdays, and a capped pursuit budget, and she wrote one line in the file: we win this on the team or we do not win it.

The preparation

Three things Kestrel did that the losing firms, by Pri Sethi's later account, did not.

One: they read the scoring matrix and built the presentation from it. Meridian published the weights in the RFP. Kestrel's deck was outlined against those weights and timed against them — twenty-two minutes on the team, thirteen on Meridian's operational constraints, eleven on the preconstruction approach, and under six on fee.

Two: Margo Deacon walked the clinic, not the site. Every firm walked the 6.2-acre site. Kestrel's general superintendent also went to the adjacent Meridian clinic — the one on the north property line that would stay open six days a week for all 565 days — and stood at its loading dock at 6:00 a.m. on a Tuesday to watch deliveries, then again at 4:30 p.m. to watch the patient drop-off queue. She counted vehicles. She timed the ambulance approach.

Three: Ray Alvarez built an answer to the risk question before it was asked. Kestrel's preconstruction team produced an independent cost model from the schematic drawings — carried at a level of detail closer to design development than schematic — and expressed the result as a range with named drivers, not a number.

The interview

Seventy-five minutes. Kestrel used sixty-one and gave the rest back.

The moment that decided it came about forty minutes in, when a member of Meridian's facilities committee asked how Kestrel would keep the clinic operating.

Margo Deacon: "Your dock takes eleven deliveries between six and eight in the morning, and your patient drop-off backs up to about nine cars around four-thirty. So we don't put our gate there. Our construction entrance goes on the east side off Halloway, which costs us about four hundred feet of haul road and is worth it. Our crane sits on the west, and it does not swing loads over your canopy — not ever, not on a Sunday. And when we have to touch your ambulance approach, you get fourteen days' written notice, it happens on a Sunday, and it is done in under four hours."

Then the risk question, which Ray had been waiting for.

Ray Alvarez: "Two things are most likely to move this number. One is structural steel escalation between the GMP and buyout — we're going to carry that explicitly on its own line rather than bury it, so you can see it. Two is the imaging equipment. Your vendor isn't selected. Whatever unit you pick sets the depressed slab, the shielding, and the electrical feed. We need that decision by a date we'll agree on, and if it comes late, it will be a change order. I'd rather tell you that today than explain it to you in eighteen months."

The score

Meridian's committee had seven voting members. The published matrix, and the totals as Pri Sethi later shared them:

Kestrel lost the qualifications category to the national builder by four points and lost fee to everyone. It won on the two categories that were worth 40 points between them — the team, and understanding the owner's constraints — by margins of seven and six.

Analysis

Kestrel did not win on price, and it did not win on company. It won 24 of 25 available points on "the proposed team" because the people in the room were the people who would run the job, and one of them had stood on the client's loading dock at six in the morning. The national builder scored a perfect 20 on qualifications and a 17 on team, and lost by twelve points overall. That is the structure of qualifications-based selection working exactly as designed: the owner is buying a group of humans for two years, not a logo.

Naming the risk was worth more than hiding it. Ray's answer about the imaging equipment is the kind of thing firms are trained not to say, on the theory that raising a problem creates one. In a room of people who have all lived through a project going wrong, it does the opposite. It converts you from a vendor making claims into a builder describing reality.

And it did not prevent the problem. Seventeen months later, Meridian's imaging vendor selected a different MRI unit after the GMP was set. The depressed slab got deeper. That became change order #14 — $186,400 of cost incurred, only $121,000 substantiable from contemporaneous records, $142,750 negotiated eight weeks later, and $43,650 Kestrel never recovered. Predicting a risk out loud buys you goodwill and a faster negotiation. It does not buy you a claim, and it is worth a small fraction of what a written directive would have been worth.

The two promises

Kestrel made two commitments in that interview that it then had to live with.

Promise one: a guaranteed maximum price by January of Year 1, at roughly 65% construction documents. That is thirty-five percent of the drawings Kestrel guaranteed a price against. The entire cost of that promise sits visibly in the GMP build-up: a $1,320,000 construction contingency and a $575,200 escalation allowance inside a $47,500,000 number. Kestrel was paid a 4.0% CM fee on the $45,120,000 subtotal — $1,804,800 — and in exchange carried the design-development risk on a third of the documents.

Promise two: the north property line. The clinic stays open six days a week. No more than four interruptions of its ambulance and patient drop-off access for the whole 565 calendar days, each with fourteen days' written notice, each on a Sunday, each under four hours. That single sentence reshaped everything downstream — the construction entrance moved to the east side, the tower crane went west, delivery windows were built around the clinic's dock schedule, and the enclosure sequence had to work without ever staging on the north line.

None of that was free. It is the part of a winning interview nobody talks about: a promise specific enough to win with is specific enough to be held to. Kestrel's site logistics plan was written in that interview room, fourteen months before anyone mobilized.

Discussion questions

1. Kestrel scored 16 out of 20 on qualifications and still won by twelve points. If you ran a firm with weaker healthcare experience than every competitor, which two categories would you attack, and how would you spend your preparation hours differently?
2. Ray named the imaging-equipment risk in the interview and the risk materialized anyway, costing Kestrel $43,650 it never recovered. Was naming it a mistake? What specific contract or process mechanism, agreed at the preconstruction stage, would have converted that prediction into protection?
3. Kestrel scored lowest of the four firms on fee and won anyway. Under what circumstances would a fee gap become decisive in a matrix that weights fee at only 10 points?
4. The north-property-line promise is a competitive advantage and an operational constraint at the same time. Estimate, in general-conditions terms, what four Sunday shutdowns with fourteen days' notice cost Kestrel across 565 days — and say whether you would have made the promise anyway.
5. Meridian shortlisted four firms from nine. What would you put in a statement of qualifications to survive that first cut, given that no price and no interview are involved at that stage?

Your turn

Take the scoring matrix above and write Kestrel's 75-minute agenda — every segment, its minute count, and who is speaking. Allocate minutes in proportion to the point weights, name the person on the team who owns each segment, and write the two-minute close verbatim. Then mark the one segment you would cut first if the panel ran fifteen minutes late, and say why that is the right thing to lose.