Engineering and architecture
Fee proposal data: should fees be in a licensed dataset?
By SourceX Editorial · Reviewed by Noah Loul ·
Short answer
Architecture fee proposal data should be excluded from a licensed dataset by default, or included only as bands with client and competitor details removed. A proposal's scope, assumptions, staffing plan and go/no-go reasoning usually carry the value; exact fees and rates add competitive exposure. Include precise fees only with counsel's review and strict use limits.
Key takeaways
- Default rule: exclude exact fees and billing rates, or band them, unless counsel approves precise figures.
- Hours by role and phase, scope assumptions and win or loss reasons hold most of a proposal's value without its prices.
- Recent fees are more sensitive than old ones, and public-sector pursuits need a separate review.
- Proposals contain staff resumes and client contacts, so personal data must be removed before any license.
What is in an A/E fee proposal record?#
An A/E fee proposal record is more than a price: it is the firm's plan for the work, its assumptions and, later, the outcome of the pursuit. Most firms keep it in pieces across a CRM such as Deltek Vantagepoint opportunities, an Excel fee build-up, the proposal document itself and email.
The pieces rarely share a key. Vantagepoint opportunity records hold the stage, outcome and competitors, while fee build-ups sit in spreadsheets saved under inconsistent file names. Linking each build-up to its opportunity number is usually the main preparation task, and it is worth doing before anyone debates whether fees stay in.
The fee is one line in that list. The hours, assumptions and outcome are what show how a firm thinks about scope and risk, and they are the parts that survive preparation intact.
- Scope narrative, phase breakdown and deliverables list.
- Hours by role and phase, and the billing rates applied to them.
- Consultant fees, reimbursable allowances and contingencies.
- Assumptions, exclusions and triggers for additional services.
- Schedule and staffing plan, often with key personnel resumes.
- Go/no-go decision, win or loss outcome, debrief notes and negotiated changes.
Why are exact fees the riskiest field?#
Exact fees are the riskiest field because they reveal pricing that competitors and clients could use against the firm, and they add little that a scope-to-effort pattern does not already show. A licensed dataset passes through a buyer's systems and may shape tools many firms use, so pricing should be handled as if it could become visible.
There is also a legal angle. Exchanging current, firm-specific pricing with competitors can raise antitrust concerns, and while licensing historical records to an AI developer is not the same as sharing prices with a rival, counsel should review whether and how any fee data leaves the firm. Client agreements may also treat negotiated fees as confidential.
Public-sector work adds another layer. Under qualifications-based selection, fees are negotiated after the firm is chosen, and negotiation records can be sensitive even when final contract values are public.
Risk table: competitive exposure vs value added#
The risk table shows that exact fees, billing rates and negotiated reductions carry the most competitive exposure for the least added value, while hours by role, scope assumptions and win or loss reasons add the most value at moderate or low exposure. Fields with high exposure and low added value are excluded; fields with low exposure and high value are kept after preparation.
| Field | Competitive exposure | Value to AI developers | Default treatment |
|---|---|---|---|
| Exact lump-sum fee | High | Low to moderate | Exclude or band |
| Billing rates by role | High | Low | Exclude |
| Hours by role and phase | Moderate | High | Include, without rates |
| Fee relative to construction cost | Moderate | Moderate | Band |
| Scope narrative and assumptions | Low once the client is removed | High | Include after preparation |
| Exclusions and additional services triggers | Low | High | Include |
| Go/no-go reasoning | Moderate | High | Include after review |
| Win or loss and stated reason | Moderate | High | Include with competitor names removed |
| Negotiated reductions | High | Moderate | Exclude or summarize |
| Client names, contacts and resumes | High | None | Remove |
Decision rule: when to exclude, band or include fees#
The decision rule is to exclude exact fees by default, band them when a buyer needs a sense of scale, and include precise figures only with counsel's review, an age cutoff and license terms that limit use. Each step down that list needs a stronger reason than the one before.
Make sure bands do not quietly re-identify projects. A band that holds only one project of a given type, sector and region is a price, not a band.
- Exclude exact fees and billing rates from any first package.
- Band fees into wide size ranges when the scope-to-effort pattern needs scale.
- Express fee relative to construction cost as a band rather than an exact ratio.
- Apply an age cutoff so recent proposals are excluded or banded more heavily than older ones.
- Include precise fees only when counsel approves and the license bars redistribution and use against the firm.
Why link proposals to what happened on the project?#
Linking a proposal to the project that followed turns a pricing document into a forecast with a known result. The proposal's hours by role and phase become the plan, and the timesheets and change orders from the delivered project become the outcome, so the gap shows where the firm's assumptions held and where they broke.
That link is often worth more than any fee figure. A proposal that assumed a single round of agency review, paired with a project record showing several resubmittals and the additional services request that followed, teaches more about scoping than a precise lump sum ever could.
Keep the link at the level of phases and assumptions, not dollars. The variance can be expressed in hours or relative to the proposed effort, which preserves the lesson while the fee stays excluded or banded.
What personal data hides in proposal records?#
Proposal records hide more personal data than most CFOs expect: key personnel resumes, client contact names and email addresses, references and sometimes billing rates tied to named staff. All of it is removed or replaced with role codes during preparation.
Laws on personal data may apply when records are licensed, including state privacy laws and, in some states, rules for businesses that sell or license personal data. Removing personal data reduces that exposure, but which laws apply is assessed deal by deal with counsel.
Illustrative: a regional A/E firm prepares its proposal history#
Illustrative: a fictional regional architecture and engineering firm keeps opportunities in Deltek Vantagepoint, fee build-ups in Excel and debrief notes in a shared folder. Its CFO is asked whether proposal history could be part of a licensed package and worries first about the firm's rates.
The CFO keeps scope assumptions, exclusions, hours by role and phase, go/no-go reasoning and win or loss reasons. Billing rates and negotiated reductions are removed, fees become wide size bands, client names become sectors, resumes are dropped and public-agency pursuits are set aside for counsel to review.
The hardest step is not the fees but the matching: many build-up spreadsheets were never saved under an opportunity number. The CFO links the pursuits that can be matched with confidence, leaves the rest out of scope, and now requires every new fee build-up to carry its Vantagepoint opportunity ID so the link exists from the start.
How SourceX handles fee fields#
SourceX handles fee fields in the Preparation and Approval steps of the SourceX five-step transaction, and the firm decides each field's treatment. Nothing is shared during the fit check, which asks only which systems hold proposal records and how far back they go.
The SourceX Evidence Packet's privacy record lists every field removed, banded or kept, and the license states the permitted use, so the firm can show its partners exactly what was licensed and on what terms.
Frequently asked questions
Do AI developers want fee data at all?
Some do, for tools that estimate effort or draft proposals, but the relationship between scope, assumptions and hours usually matters more than the price itself. Exact fees are rarely essential, which is one reason excluding or banding them costs a firm little.
Can bands still reveal our pricing?
They can if they are narrow or contain very few projects. Use wide bands, check that each band holds several comparable projects, and combine banding with removal of client names, locations and any dates precise enough to identify a specific pursuit.
Are fees on public contracts safe to include because they are already public?
Public availability reduces confidentiality concerns but does not remove them. Negotiation records, rates and internal build-ups usually are not public, and agency contracts can carry their own records terms. Review public pursuits separately with counsel before including them.
Could a licensed dataset with our fees end up helping competitors?
That depends on the license. Terms can restrict permitted use, bar redistribution and limit who may access the data, but a model trained on many firms' records may still be offered to the market. Set permitted use deliberately and exclude anything you would not want generalized.
Should we tell clients that proposal history was licensed?
Check client agreements first, because some require notice or consent for any use of project information. Where client names are removed and no confidentiality terms apply, notice may not be required, but some firms tell key clients anyway as a matter of relationship.
Related resources
See if your company qualifies
A short company assessment. No data uploads are needed.