Backlog and book-to-bill in engineering firm due diligence: what buyers test and what it proves

How do buyers test backlog in an engineering firm?

Buyers test backlog by asking how much of it is signed, how much of it converts to revenue and how much depends on a few people. Backlog is the contracted work not yet performed; book-to-bill compares new work booked with revenue billed over a period. A ratio above one suggests the pipeline is growing, and below one suggests it is shrinking, but neither number means much until the quality of the contracts behind it is known.

For a sell-side advisor, the useful point is that the evidence buyers ask for is also the firm's best set of work records. Proposals, project files, change orders and QA/QC reviews prove backlog quality in diligence, and the same records are what AI developers want when they train and evaluate agents on real professional work.

This explainer covers the tests buyers apply, how to prepare the evidence, and how a licensing conversation fits next to a sale. This explainer does not quote market multiples or benchmark ratios; use your own deal data for those.

What do backlog and book-to-bill actually measure?

Define each metric in the CIM the way the firm actually calculates it. If your definition differs from the buyer's, the numbers will be re-cut during diligence and trust drops.

What makes backlog high quality or low quality?

The 5-point backlog quality screen asks five questions of every large line item.

1. Is it contracted? Executed agreement and notice to proceed, or only a verbal award or a shortlist.
2. Is it funded? Public owners may have appropriations still pending; private owners may have financing contingencies.
3. What is the contract type? Lump sum, time and materials, cost plus or unit price carry different margin and risk.
4. Who delivers it? Named principals and project managers who might leave or retire after a sale.
5. Can the owner cancel? Termination for convenience clauses, option years and task-order structures reduce certainty.

Principal dependence is the item advisors underestimate. If the three people who hold the owner relationships are all approaching retirement, a buyer will discount even a strong number.

How should an advisor build the backlog evidence pack?

Assemble the pack before the first management meeting so numbers and documents tie out.

• Backlog schedule by project: client, contract type, award date, contract value, billed to date, remaining value, expected burn dates
• Three to five years of year-end backlog and quarterly book-to-bill
• Win-loss log for proposals, with reason codes
• Contract copies or summaries for the largest items, including termination and change-order terms
• Staffing plan mapping principals and project managers to major projects
• Margin by project at award and at completion
• Slippage and write-down history, with root causes
• QA/QC and peer-review sign-offs for projects reaching completion

Each row traces back to a document inside a system: the CRM, the proposal library, the project management platform, the accounting system, the document control system. That trace is what makes the pack credible. It is also the inventory of where the firm's work records live.

Where do licensable work records sit inside the same evidence?

An engineering firm's proof of backlog quality is built from its own working records. Several of them overlap with what AI developers seek for agent training and evaluation.

Three limits matter. Drawings, calculations and reports prepared for a client usually belong to that client, or are restricted by contract, so they are excluded unless the client consents. Documents may carry stamps, licensure or security restrictions, especially on public or critical infrastructure work. And employee personal data is handled under redaction rules agreed with the company before any work begins.

The firm's own records qualify when it holds the rights. Companies need 50+ full-time employees at peak (contractors excluded), several years of documented operations, rights to license the data and an authorized sponsor; the who qualifies page has the detail.

How does this fit into the sell-side process?

Raise it at the same moment you collect the backlog pack, because the system map is already on the table.

1. During preparation, list each system holding proposal, project and QA/QC history and the years it covers.
2. Ask the firm's controller or IT lead who can export each system, and whether any platform is scheduled for retirement. Preserve a complete export first.
3. Agree with the owner whether a license should close before, during or after the sale. An exclusive AI-training term can continue after closing, so the acquirer should hear about it early.
4. Screen the firm with the company fit checker, a preliminary, non-binding check that needs no contact details.
5. If the owner is interested, introduce the firm. The company completes an inventory, price and terms are agreed, buyers review, and nothing is binding until the company signs.

The same diligence discipline applies to technology claims; the guide to AI washing in due diligence shows how to test them, and the data quality due diligence checklist helps with a target's own records. For recurring-revenue targets, see how to build an ARR bridge. Acquirers' internal teams can read about referral opportunities for corporate development teams.

What do rewards look like for an advisor?

Partners earn 25% of the eligible platform fees SourceX actually collects from the referred company's licensing deals, capped at $100,000 per referred company. The reward is paid only after the buyer pays and SourceX receives its fee. A lead, a meeting or a signed agreement alone does not trigger payment, and no reward is guaranteed. The reward is a share of SourceX's fee and is never deducted from what the company receives. Check your engagement letters and licensing rules before accepting any referral fee; the referral opportunities for M&A advisors page covers the advisor view in more detail.

When is this the wrong conversation?

• The records are mostly client drawings and calculations with no consent.
• A court, trustee or assignee controls the assets and has not been involved.
• Archives were deleted or systems were cancelled without export.
• The firm already licensed the same data for AI training.
• The owner will not consider an exclusive license.

Next step

If you advise engineering or architecture owners, register as a partner and run the first firm through the fit checker while you assemble its backlog pack. A company can also apply directly at sourcex.si/apply.