How to Score and Shortlist Acquisition Targets You Can Defend

Target selection is where deal outcomes are decided, long before integration begins. Baruch Lev and Feng Gu analysed 40,000 acquisitions worldwide over the past 40 years and found that 70 to 75 percent failed to achieve their stated objectives, such as enhancing post-acquisition sales growth, cost savings, or maintaining the buyer's share price.^[1] Their finding sits inside a broader pattern: study after study puts the failure rate of mergers and acquisitions somewhere between 70 and 90 percent.^[2] Worse, Lev and Gu's data show a reverse learning curve, with failure rates rising over time rather than falling as organisations accumulate deal experience.^[1]

Part of the explanation sits at the very top of the funnel. When screening is unstructured, deal flow drifts toward whatever arrives opportunistically: banker-led pitches, broker introductions, and targets that happen to be for sale. Lev and Gu describe executives yielding to investor pressure and the advice of commission-hungry investment bankers, then buying a strategically misfit target.^[1] A scorecard is the countermeasure. Used properly, it is not a ranking tool but a decision record that concentrates scarce diligence capacity on the targets most likely to create value, and it does so against criteria agreed before anyone fell in love with a deal.

• It forces the acquisition objective to be written down as measurable criteria before sourcing starts.
• It makes trade-offs explicit, so a strong financial profile cannot quietly compensate for weak strategic fit.
• It creates an auditable artifact: months later, the board can see not just which targets were shortlisted but why.
• It disciplines deal flow, so opportunistic proposals must clear the same bar as proactive ones.

The payoff for this discipline is measurable. McKinsey's analysis of the Global 2,000 found that programmatic acquirers, those making more than two small to midsized deals annually over ten years, delivered a median excess total shareholder return of 3.9 percent over the past decade, up from 2.9 percent in the 2010s, outperforming every other M&A strategy including organic growth.^[3] Repeated, criteria-led screening is exactly the habit that separates those acquirers from one-off dealmakers, and it is the mindset tools like Decisity M&A & Corporate Finance are built to support.

Institutional acquirers converge on a small set of screening dimensions, but the content and weighting of each depends entirely on why the deal exists. An expansion acquisition, a capability acquisition, and a consolidation acquisition fail for different reasons, so they must be screened with different logic. The dimensions below form the common skeleton; the thresholds inside them are where the acquisition objective does its work.

Two design rules make the difference between a scorecard and a spreadsheet. First, weights must follow the acquisition objective. A consolidation deal may justifiably weight cost synergies and market share heavily, while a capability acquisition should weight talent, technology, and retention, even at the cost of near-term margins. Second, define exclusion criteria alongside qualifying criteria. Regulatory exposure is the clearest example: public companies must generally file a Form 8-K disclosing a material cybersecurity incident within four business days of determining materiality, so an unresolved, undisclosed incident at a target is a jurisdiction-specific red flag worth screening for explicitly.^[4] Customer concentration and unresolved litigation belong in the same exclusion tier: they are the conditions under which no score on the qualifying criteria should matter.

A score without a source is an opinion wearing a number. The traceability discipline is simple to state and demanding to keep: every score on the scorecard points to one named source, and the pointer records the page, table, or cell the figure came from and the date it was retrieved. This is what allows a reviewer, an investment committee member, or a due diligence team to walk from any score back to its evidence without redoing the analysis, and it is the same principle behind auditable strategy decks with traceable sources.

The claim-to-source map

Think of the scorecard as two linked layers: the scoring layer that the committee reads, and a claim-to-source map underneath it that a reviewer audits. The map is what survives scrutiny. When a board member asks why a target scored three on customer concentration, the answer is a document name, a page, and a date, not a shrug. Practitioners who review AI-drafted analysis make the same distinction between attributable claims and source-flavoured prose, and the habit applies equally to human-drafted scorecards. Fact-check a strategy deck

• Source identity: the named document or dataset, not a vague reference to management estimates.
• Location: the page, table, or cell the figure came from.
• Retrieval date: when the source was accessed, because targets change and filings get amended.
• Status: whether the source is primary (audited accounts, filings) or secondary (press, broker notes).

Just as important is recording what could not be verified. A scorecard that marks a criterion as unverified, with a note on what evidence would settle it, is more credible under challenge than one that claims completeness. Gaps acknowledged at shortlisting become diligence workstreams; gaps discovered at integration become write-downs.

The scorecard earns its value as a decision record only if it is version-locked at the moment the shortlist is agreed, before any target becomes a favorite. From that point on, the criteria, weights, and thresholds are a fixed baseline against which every later decision is measured. The record's defensibility rests on this freeze: it proves the yardstick existed before the winner was known.

Revising weights or thresholds after the winner emerges is the first thing a board or investment committee challenges, and rightly so. A baseline adjusted to fit the outcome is not a decision record; it is a rationalization with formatting. The pressure to adjust is rarely cynical. Once a team has spent weeks on a target, the temptation to re-weight a criterion that penalises it is human. The freeze exists precisely because the temptation is predictable.

1. Lock the scorecard and record its version identifier when the shortlist is formally agreed.
2. Route any proposed change through formal review, with the justification, the reviewer, and the date documented.
3. Re-score affected targets under the revised baseline and record both versions, never overwrite the original.
4. Treat ad hoc adjustment as a governance breach, not a workflow preference.

Finally, freeze the inputs, not just the output. The source documents, the criteria definitions, and the original brief must all be archived together with the signed scorecard. A locked summary pointing at sources that have since moved or changed is a record that cannot be reconstructed, which defeats the purpose. Lev and Gu's finding that executives frequently buy strategically misfit targets under pressure^[1] is the strongest argument for a baseline that cannot quietly move.

The artifact itself should fit on one page per target. Anything longer stops being used, and a record that is complete but unusable protects no one. The structure below has one row per criterion and one source per row, with a scorer and a sign-off so that every judgement is attributable to a person.

The rows above are worked examples of the format, not recommendations for any specific weighting. Note what they encode: the strategic-fit row cites the documents the fit judgement rests on, and the financial row cites a specific statement and page. One page per target keeps the record usable in the meeting where it matters, and the full source set travels behind it.

Traceability tells the board what the scores rest on; governance tells them who stood behind the process. The approval chain runs in two stages. First, the screening criteria, weights, and exclusion rules are approved by the board or investment committee before sourcing begins, so the yardstick is sanctioned before any target exists to benefit from it. Second, each shortlist decision carries named reviewers and a sign-off, so the decision to advance or drop a target is attributable to specific people on a specific date.

1. Board or IC approves the criteria set, weights, and deal-breakers before sourcing starts.
2. Deal team scores targets against the frozen criteria, one named scorer per criterion.
3. Named reviewers check scores against sources, and the shortlist goes to sign-off with the scorecard attached.
4. Any deviation from the approved criteria requires explicit written justification and re-review by the approving body.

The second point deserves emphasis: review that cannot be attributed to a person is indistinguishable from no review. Anonymous or collective sign-offs dissolve exactly when they are needed most, under challenge. This is the same reasoning behind tiered human review protocols for AI-assisted board decks, where analyst, engagement owner, and independent signatory roles carry distinct, named responsibilities.

Deviations are where governance earns its keep. Opportunistic deals bypass discipline by arriving with a story that excuses them from the criteria; the requirement for explicit justification and re-review is what stops that. McKinsey's finding that the best-performing acquirers run programmatic, actively governed portfolio strategies rather than opportunistic ones^[3] is, at bottom, a governance result: the process, not the anecdote, decides which deals proceed.

The real test of a scorecard is whether it works when nobody involved is in the room. Hand the frozen scorecard and its complete source set to a colleague who was not part of the deal team, and ask them to reconstruct the recommendation and its load-bearing scores within a short, fixed time. If they can, the record works. If they cannot, every failure mode maps to a missing artifact.

• Dead links and changed pages point to sources that were never frozen with retrieval dates and archived copies.
• Scores that cannot be traced point to a missing claim-to-source map between the scoring layer and the evidence.
• Unexplainable weights point to criteria that were never approved, or approved after the fact.
• No attributable sign-off points to a governance chain that existed in practice but not on paper.

Reconstruction is not bureaucratic nostalgia; it is how organisations learn. Lev and Gu's data show M&A failure rates have risen over time rather than fallen,^[1] which means most acquirers are not learning from their own record, usually because no usable record exists. The same frozen-record principle applies to AI-assisted analysis: our guide on reproducing an AI strategy analysis six months later shows that reproducibility depends on the record kept at sign-off, not on the tool that produced the draft.

This is where a platform built for traceability changes the economics of the discipline. Decisity is an AI-native strategy consulting platform in which every conclusion is traceable to its source, so the scorecard, its claim-to-source map, and the frozen inputs are produced as part of the work rather than assembled after it. For teams screening targets across markets and languages, that means the record a board sees is the record a reviewer can reconstruct. See how the three-step workflow carries a brief from document ingestion to a board-ready, source-linked deliverable.

Related reading in this series: Making a Build-Buy-Partner Recommendation the Board Can Trust and Making an Acquisition Valuation Defensible: The Price Bridge.

• Fortune: Analysing 40,000 M&A deals over 40 years — why 70–75% fail
• Harvard Business Review: The New M&A Playbook
• McKinsey & Company: Top M&A trends — blueprint for success in the next wave of deals
• US SEC: Fact sheet on enhanced disclosures in M&A transactions