Bid Analysis & Scoring Model

A dispositions manager comparing buyer bids is rarely choosing on headline price alone. Two offers at the same number can differ on deposits, financing, inspections, board approvals, and how soon the buyer can actually close. This page describes a bid analysis and scoring model that parses submitted offer packs, scores each bid on price, conditionality, funding certainty, and timeline, and returns a ranked summary the asset manager uses to choose. The model ranks. It does not award.

Related work on the same disposal path: Asset Valuation Model & Comparable Analysis, Buyer Targeting & Mandate Matching, and Data Room Preparation Agent.

The model treats each complete bid pack as a structured record. It extracts the offered consideration, the condition stack, evidence of funds or financing, and the proposed close path. It then scores those dimensions against the sale mandate and the asset's documented constraints, and it produces a ranked list with the drivers of each rank visible.

The unit of work is one bid on one asset in one process. A round of first-round offers, a best-and-final set, or a small bilateral pack all follow the same rule: only documents that belong to that round and that asset are scored together. Mixing rounds, mixing assets, or scoring an unsigned term sheet as if it were an executed offer produces a ranking that looks precise and is not usable.

The ranking is a decision aid. The asset manager still selects the counterparty, still negotiates, and still signs. Nothing in the scoring output executes an award, sends an acceptance, or closes other bidders.

Scoring starts only when a bid pack is present. Typical inputs include the offer or purchase agreement, mark-ups to the seller's form if used, proof of funds or financing commitment letters, deposit mechanics, proposed conditions and drop-dead dates, and any buyer comments on title, environmental, or operational diligence already disclosed in the data room.

The sale mandate sits beside the packs. The model needs the asking or reserve framework the team is using, the preferred close window, known deal-breakers (for example vacant possession, assumed debt, or a required entity structure), and any scoring weights the dispositions committee has already set. If weights are not set, the model still scores each dimension separately and ranks with an explicit default weighting that the manager can override before selection.

Identity of the bid matters as much as the numbers. Each pack must be attributable to a named bidder, a round, and a timestamp. Duplicate submissions from the same buyer in the same round are collapsed to the latest complete pack unless the manager flags an earlier version as the operative one.

When bid documents are missing, incomplete beyond a usable extract, or not attributable to a bidder, the model returns empty output for that bid. It does not invent a price from a cover email, does not infer a financing letter from a verbal claim, and does not fill a condition list from a prior round unless that prior pack is explicitly in scope. An empty result is the correct state, not a failure to "guess well."

Price is scored relative to the mandate and, where a valuation pack exists, relative to the current valuation view of the asset. Headline consideration is not enough. The model reads net proceeds effects that are written into the bid: seller credits, assumed liabilities, deferred purchase price, earn-outs, and deposit timing. A higher sticker with a large seller credit can rank below a cleaner number. Price scoring stays inside what the documents state. It does not revalue the asset; that work lives in Asset Valuation Model & Comparable Analysis.

Conditionality is scored by counting and weighting conditions that can still kill or delay the deal after selection. Financing conditions, board or investment-committee approvals, further diligence, planning or licensing dependencies, and third-party consents all increase conditionality. Waived or already-satisfied conditions reduce it. The model distinguishes conditions that are binary walk-away rights from conditions that only adjust price or timing. A bid that looks unconditional on page one but restores a financing out in a schedule is scored on the schedule.

Funding certainty is scored from the evidence in the pack, not from the buyer's reputation in the abstract. Cash with verifiable proof of funds ranks differently from a highly confident but uncommitted financing plan. Commitment letters, allocation memos, and deposit already in escrow are treated as stronger than expressions of interest from a lender. If the pack contains no funding evidence, that dimension scores at the floor and is flagged. The model does not treat a well-known sponsor name as a substitute for documents.

Timeline is scored against the mandated close window and against internal constraints such as loan maturity, fund life, or a vacant-possession date. Proposed exclusivity, long diligence, delayed deposit, and open-ended extension rights all pull the timeline score down even when price is strong. A fast close that depends on an unmet condition is not scored as fast.

The four dimension scores are combined into a composite only after each is visible on its own. The ranking uses that composite plus hard constraints from the mandate (for example: no award below a stated floor, or no bid that requires a structure the seller cannot deliver). A bid that fails a hard constraint is still listed, marked ineligible, and excluded from the recommended rank order. Hiding it would make the manager's file incomplete.

The primary output is a ranked bid summary for the round. Each row identifies the bidder, the extracted price and net-proceeds view, the conditionality score, the funding-certainty score, the timeline score, the composite, eligibility against hard constraints, and a short list of drivers (what lifted or lowered the rank).

The summary is meant to be read in a selection meeting. It should answer: which bids are eligible, where they differ on risk of fail-to-close, and which gaps would need to be closed in negotiation if a lower-ranked bid is still the commercial choice. It is not a recommendation letter and it does not contain an award decision.

Supporting extracts sit behind each row: cited clauses for conditions, cited pages for funding evidence, and the close dates as written. If an extract cannot be tied to a page or clause, that field stays empty rather than paraphrased from memory of the process.

The asset manager uses the ranking to select, to invite a second round, or to negotiate with one or more bidders. Those actions stay with the manager and counsel. The model does not auto-award, does not notify bidders, and does not lock a preferred bidder status. If the manager selects a bid that is not rank one, that is a valid outcome. The file should show the ranking that was on the table and the human decision that followed.

Downstream, a selected bid still has to survive documentation and closing. Scoring does not replace legal review of mark-ups, KYC, or the data-room trail that Data Room Preparation Agent assembled. Buyer origination quality from Buyer Targeting & Mandate Matching also does not substitute for the pack in front of you; a well-matched mandate can still submit a weak or highly conditional bid.

Empty output is required in defined cases. No bid packs in the round: empty ranking. A pack with no extractable offer terms: empty row for that bidder, not a placeholder score. Conflicting executed documents in the same pack with no manager instruction on which governs: empty for the conflicted fields, with the conflict flagged. Scoring weights or mandate missing when the team has required them before any ranking is shown: empty ranking until those inputs exist.

Partial rounds are allowed only as partial tables. If three of five invited bidders have submitted complete packs, the model ranks those three and lists the other two as not submitted. It does not impute their likely bid from earlier marketing conversations.

The dispositions manager, asset manager, or investment committee remains the decision-maker. The model can be wrong on a clause, can under-weight a relationship or a portfolio trade the documents do not capture, and can over-weight a clean but low price. Those are reasons for a human to select against the rank, not reasons for the model to award on its own.