# RFP Reporting Bundle

Build and refresh a one-page bid reporting dashboard for sales leadership with the 7 numbers a CRO asks for (pipeline influenced, revenue leakage, go/no-go accuracy, SME bottleneck tax, content use, requirement gaps, churn and contraction from RFPs) plus a war room of the largest bids in flight. Use when asked to build a bid reporting dashboard, measure the 7 numbers, report bid team metrics to sales leadership or a CRO, set up a weekly bid dashboard, or turn CRM and RFP platform data into revenue reporting. Also writes the one-page CRO pre-read for the forecast call.

This file is the Agent Skills standard package flattened for clients that take instructions rather than a Skill folder. Keep workflow files in Instructions. Upload business source documents and optional assets as Knowledge only when needed.

## SKILL.md

# RFP Reporting Bundle

Builds one page that reports the bid team's work in sales leadership's units: seven numbers, each for this period against the last, and a war room of the largest bids in flight. It sits beside the volume, turnaround and cost numbers the team already reports.

## Pick the workflow

- **Bid reporting dashboard** (default): follow the steps below.
- **CRO one-pager**, a written pre-read for the CRO before the forecast call: read [CRO one-pager](references/rfp-sales-report.md). It can reuse the setup, the connection map and the numbers from a dashboard run.

## 1. Setup

Look for a saved **bid reporting setup** first: a Notion page, a Doc artifact, or `bid-reporting-setup.md` in a connected folder or project files. If there is none, look for an `rfp-sales-report` profile (see the CRO one-pager) and reuse what it already answers.

Ask only for what is still missing: company name, currency, who reads the page (CRO, VP Sales, revenue leadership), the forecast call day, the fiscal calendar, the deal-size bands to cut by, and the forecast categories that make a deal must-win (default: commit and most likely).

Show the setup, ask where to save it (Notion page, Doc artifact, Word document, SharePoint, or a markdown file in a connected folder), and save it only after the user approves. Date it.

## 2. Connections

Read [Connections](references/connections.md) and follow it:

1. Check what is connected: CRM, RFP platform, content library, customer success or renewals tool, and wherever SME tasks live.
2. Map every data point each number needs to a source, using the field-to-source map in that file.
3. For a missing source, search the connector registry and suggest a connector.
4. If there is still no connector, ask for an export in the tab layout of `assets/bid-reporting-dashboard.xlsx`.
5. Show the map and confirm it with the user before pulling anything. Save it with the setup.

## 3. Definitions

Ask one round of questions, up to four (AskUserQuestion in Claude, one short message elsewhere). Skip any the setup already answers:

1. **Bid-touched deal:** what links a deal to a bid? A CRM field or checkbox, a link from the RFP platform project to the opportunity, or a match on account name within a date window. Recommend the CRM field when one exists.
2. **Period:** this quarter against last quarter, this month against last month, or another span the forecast call uses.
3. **Go/no-go calls and overrides:** where are the call and the person who overrode it recorded? RFP platform, a CRM field, a spreadsheet, or nowhere yet.
4. **Leak reasons:** tag leaks by the four hand-offs (sales to bid, bid to legal, negotiation, delivery), or by the team's own list?

If nobody is there to answer, use the recommended option, state the assumption in the footer, and continue. Save the answers into the setup. Ask the user to confirm the definitions with RevOps, so the bid team's pipeline number matches theirs.

## 4. Compute

Pull records into files and compute in code (Python, or the spreadsheet's formulas). Never work a number out in prose. Take one number at a time and read its reference before computing it:

| Tile | Reference |
| --- | --- |
| Pipeline influenced | [pipeline-influenced.md](references/pipeline-influenced.md) |
| Revenue leakage | [revenue-leakage.md](references/revenue-leakage.md) |
| Go/no-go accuracy | [go-no-go-accuracy.md](references/go-no-go-accuracy.md) |
| SME bottleneck tax | [sme-bottleneck.md](references/sme-bottleneck.md) |
| Content use | [content-use.md](references/content-use.md) |
| Requirement gaps | [requirement-gaps.md](references/requirement-gaps.md) |
| Churn and contraction from RFPs | [churn-contraction.md](references/churn-contraction.md) |
| War room | [war-room.md](references/war-room.md) |

For every number:

- Compute this period and last period with the same definition, then the change.
- Join bids to CRM deals by the agreed method. Report the match rate and keep the unmatched records in a working file.
- When a number cannot be computed, show **Not tracked** and the field to start capturing, as its reference says. Never show zero for missing data.
- Say beside the number when the sample is small (fewer than 10 decided bids).

## 5. Output

Read [Outputs](references/outputs.md).

- **In Claude with Artifacts:** publish one page laid out like the deck's dashboard slide. Eight tiles (the seven numbers and the war room), each with its value, the change against last period and a one-line note. Below them, the war-room table of must-win deals. A footer with the definitions, sources, join match rate, period and refresh date.
- **In any other client, or when the user asks for a spreadsheet:** fill `assets/bid-reporting-dashboard.xlsx` with the pulled data (delete the example rows first) and hand the file back.

## 6. Weekly refresh

Ask whether to refresh the page on a schedule. Default: weekly, the working day before the forecast call.

- **In Claude:** create a scheduled task that re-pulls the same sources with the saved setup and map, recomputes every number, and republishes the same Artifact so the link does not change.
- **Elsewhere:** give the user the steps to rerun the export and refill the spreadsheet.

Never post, email, share or write back to a CRM or any other system without the user's approval for that action.

## 7. The written pre-read

When the user wants a written page for the CRO before the forecast call, run the [CRO one-pager](references/rfp-sales-report.md) with the numbers from this run.

## Evidence rules

- Compute every figure in code from data pulled or supplied in this run. Never estimate, extrapolate or invent a value, a benchmark or a target.
- When a figure cannot be computed, show `Not tracked` and say what to start capturing. Never fill the gap with zero or an industry figure.
- Name the period, the definitions and the source of every figure, and say how many records could not be matched.
- Keep example rows, such as the fictional rows in a reference spreadsheet, out of every real report.
- Keep account names as the source has them. Anonymize them only when the user asks.
- Send, post, email or write back nothing without the user's approval.

## Included resources

- `assets/bid-reporting-dashboard.xlsx`: the reference spreadsheet. Its tabs are the export layout (Deals, Bids, Leaks, Renewals), and its Dashboard and War room tabs compute every tile with formulas. The example rows are fictional.
- `references/`: one file per number, plus connections, outputs and the CRO one-pager.

## Churn Contraction

Source: `references/churn-contraction.md`

# Churn and contraction from RFPs

Tile: **Churn from RFPs**. Value: ARR lost or contracted after renewal-cycle RFPs that closed this period, as a share of the ARR those RFPs covered. Change: points against last period. Note: ARR at risk on renewal-cycle RFPs due in the next six months.

## What it tells a CRO

Winning is half the job. Existing customers send RFPs too: security questionnaires, DDQ reviews and re-tenders. The next questionnaire can decide whether a customer stays, so this number shows the revenue the bid team protects after the first sale.

## Fields you need

| Field | Usually in | Spreadsheet column |
| --- | --- | --- |
| Account and ARR | CRM or the renewals tool | Renewals: Account, ARR |
| Renewal-cycle RFP type | Customer success tool, RFP platform | Renewals: RFP type |
| Due date | RFP platform or the customer's calendar | Renewals: Due |
| Status | RFP platform | Renewals: Status |
| Outcome (retained, contracted, churned) and ARR after | CRM renewal opportunity or the renewals tool | Renewals: Outcome, ARR after |
| Closed date | CRM | Renewals: Closed |

## Formula

Map each customer's renewal cycle first: the quarterly security questionnaire, the yearly DDQ review, the re-tender in year three. Then:

- **ARR lost** on a renewal-cycle RFP = full ARR if the customer churned, or ARR − ARR after if it contracted.
- **Churn from RFPs** = ARR lost on renewal-cycle RFPs closed in the period ÷ ARR on all renewal-cycle RFPs closed in the period. This is the tile value.
- **ARR at risk** = ARR on renewal-cycle RFPs due in the six months after the period ends that have not closed. Six months is the default look-ahead; change it in the setup.
- Show each customer's cycle on the war room page when its ARR is above the war-room threshold.

Change against last period: this period's rate − last period's rate, in points.

## Cuts

- RFP type: security questionnaire, DDQ review, re-tender.
- Outcome: churned against contracted.
- Segment and customer size.
- Due date, by month, for the look-ahead.

## What good looks like

The bid team works with customer success to track data, quotes and results through the life of the account. There is a check-in in the first six months and a renewal playbook that starts before the questionnaire arrives. The team knows six months ahead what is coming up for renewal.

## The trap

Incumbency is not a win theme. A re-tender answered as if the renewal were safe gets scored like any other bid. Treat every renewal-cycle RFP as a new bid, with the proof gathered since the first award.

## Not tracked

- **Renewal RFPs not recorded:** show Not tracked. Start logging every security questionnaire, DDQ review and re-tender from existing customers with its account, ARR and due date.
- **Outcomes recorded without ARR after:** show churned ARR only, and Not tracked for contraction.
- **No due dates for upcoming RFPs:** show the rate for the period, and Not tracked for ARR at risk until due dates are logged.

## Connections

Source: `references/connections.md`

# Connections

Find where each number's data lives before pulling anything. Every number on the page needs a source the user has confirmed.

## 1. Check what is connected

Check silently first (in Claude: the listed MCP tools, ListConnectors or ToolSearch). Look for:

| Source | What it gives the dashboard |
| --- | --- |
| CRM | Deals: amount, status, stage, created and close dates, segment, forecast category, owner, last activity, buyer contacts, loss reasons, close-date history |
| RFP platform (AutoRFP.ai or the one the team uses) | Bids: deal link, type, received, due and submitted dates, go/no-go call and override, outcome, status, owner, SME tasks |
| Content library | Which library items went into which submitted bid |
| Customer success or renewals tool | Renewal-cycle RFPs: account, ARR, type, due date, outcome |
| Wherever SME tasks live | Assigned and answered dates for product, legal and security questions, when the RFP platform does not hold them |

Read field names from the connected systems. Do not assume a field exists because it usually does.

## 2. Map every data point to a source

Fill this map with the user. One row per data point, with the system, the object and field, and how the data arrives.

| Number | Data point | System | Object and field | Arrives by |
| --- | --- | --- | --- | --- |
| Pipeline influenced | Deal amount, status, created and close dates, size band | CRM | | connector / export / missing |
| Pipeline influenced | Link from bid to deal | RFP platform or CRM | | |
| Pipeline influenced | Bid submitted date | RFP platform | | |
| Revenue leakage | Lost or delayed, days delayed | CRM | | |
| Revenue leakage | Hand-off reason | CRM or a leak log | | |
| Go/no-go accuracy | Call, overridden by, outcome, outcome date | RFP platform, CRM or a spreadsheet | | |
| SME bottleneck tax | Days waiting by team (product, legal, security) | RFP platform or task tool | | |
| Content use | Library items used per bid | Content library or RFP platform | | |
| Requirement gaps | Gap or objection on lost and open bids | CRM loss reasons, win/loss reviews, RFP requirements | | |
| Churn from RFPs | ARR, RFP type, due date, outcome, ARR after | Renewals tool or CRM | | |
| War room | Stage, forecast category, owner, last activity, buyer contacts, needs an RFP | CRM | | |
| War room | Readiness, main risk, the ask on open bid deals | The bid team's weekly review | | |
| War room | Due date of the live bid | RFP platform | | |

A data point with no source is a number that will show **Not tracked**. List those numbers for the user now, with the field each one needs.

## 3. When a source is missing

1. Search the connector registry for the system by name (in Claude: SearchMcpRegistry), and offer the match with SuggestConnectors.
2. Connect nothing without the user's approval.
3. If no connector exists, or the user prefers files, ask for an export (section 4).

## 4. Export fallback

Ask for one file per tab of `assets/bid-reporting-dashboard.xlsx`, as CSV or XLSX: one row per record, dates as YYYY-MM-DD, amounts in one currency with no symbols. Columns in this order:

**Deals** (CRM, plus the weekly review): Deal ID, Account, Deal name, Segment, Size band, Amount, Status (Open, Won, Lost), Created, Close date, Forecast category (Commit, Most likely, Best case, Pipeline, Closed), Needs an RFP (Y or N), Owner, Stage, Last activity, Buyer contacts, Readiness (On track, Watch, At risk), Main risk, The ask.

**Bids** (RFP platform): Bid ID, Deal ID, Account, Bid type, Received, Due, Submitted, Go/no-go call (Go or No-go), Overridden by, Outcome (Won, Lost, Pending, Not submitted), Outcome date, SME days: product, SME days: legal, SME days: security, Library items used (separated by semicolons), Gap or objection, Status, Notes, Owner.

**Leaks** (CRM or the bid team's log): Deal ID, Account, Value, Leak type (Lost or Delayed), Days delayed, Hand-off reason, Date recorded, What happened.

**Renewals** (renewals tool or CRM): Account, ARR, RFP type (Security questionnaire, DDQ review, Re-tender), Due, Status, Outcome (Retained, Contracted, Churned, Pending), ARR after, Closed, Owner.

The spreadsheet adds the joined and computed columns (bid value, bid touched, cycle days, next bid due, ARR lost) itself. A missing column is fine: the numbers that need it show Not tracked.

## 5. Join bids to deals

- Use the method agreed in step 3 of SKILL.md: the CRM deal ID on the bid, a CRM checkbox, or a match on account name with the bid received between the deal's created and close dates.
- **Match rate** = bids matched to a deal ÷ all bids in the pull. Show it in the footer.
- Keep unmatched bids in a working file (`unmatched-bids.csv`) and offer to review them with the user. Never guess a match.

## 6. Confirm the map

Show the filled map, the numbers that will show Not tracked and the join method. Pull data only after the user confirms. Save the map with the setup, so the weekly refresh uses the same sources.

## Content Use

Source: `references/content-use.md`

# Content use

Tile: **Content use**. Value: win rate on bids that used library content ÷ win rate on bids that did not (shown as a multiple, such as 1.5×). Change: against last period. Note: the two win rates and the library item behind the most won dollars.

## What it tells a CRO

A content library is a cost line until it shows which answers help close deals. Joining library use to bid outcomes turns the library into a list of the answers that carry revenue, and the ones nobody uses.

## Fields you need

| Field | Usually in | Spreadsheet column |
| --- | --- | --- |
| Library items used in each submitted bid | Content library usage report, or the RFP platform's answer history | Bids: Library items used |
| Bid outcome and outcome date | RFP platform or CRM | Bids: Outcome, Outcome date |
| Deal value | CRM | Bids: Value (joined from Deals) |
| Item owner and last review date | Content library | Optional, for the action list |

List library items in one cell per bid, separated by semicolons, with the same item names every time.

## Formula

Count decided bids (won or lost) with an outcome date in the period.

1. **Tag** each library item used in a submitted bid.
2. **Join** it to the bid outcome.
3. **Win rate with an item** = won bids that used the item ÷ decided bids that used it.
4. **Win rate without it** = won bids that did not use the item ÷ decided bids that did not use it.
5. **Rank** items by won dollars: the sum of deal value on won bids that used the item.

- **Tile value** = win rate on bids with any library content ÷ win rate on bids with none.
- Show the per-item table (item, uses, won dollars, win rate with, win rate without) beside the tile or in the footer.

This is a correlation. A strong item may sit in bids that were going to close anyway. Say so in the footer, and say beside any item used in fewer than 10 decided bids that the sample is small.

## Cuts

- Item, ranked by won dollars (the main cut).
- Library section or content type (security, case studies, implementation, pricing, legal terms).
- Segment and bid type.
- Items with no uses in the period, as a list to review or retire.

## What good looks like

The team ties content to the bids it won and the ones it lost. Product and marketing help keep the items that move revenue current, and the library is shared across teams instead of kept in silos.

## The trap

Content volume is a cost. Every item in the library needs an owner and a review, so each one added makes the library more expensive to keep current. Use the ranking to cut items nobody uses as well as to promote the ones that work.

## Not tracked

- **No record of which items went into which bid:** show Not tracked. Start exporting the library usage report per bid, or tag answers with their library item when they are inserted.
- **No outcomes on bids:** show usage counts only, and Not tracked for win rates until outcomes are recorded.
- **Every bid used library content (or none did):** there is nothing to compare against. Show the per-item table and say the tile comparison is not possible this period.

## Go No Go Accuracy

Source: `references/go-no-go-accuracy.md`

# Go/no-go accuracy

Tile: **Go/no-go accuracy**. Value: share of calls that proved right this period. Change: points against last period. Note: win rate on bids the team backed against win rate on bids it advised against that were overridden.

## What it tells a CRO

Qualification calls are a forecast. Every no-go that was overridden is a bet, and this number shows whether those bets paid. It also shows the bids the team said no to, and the time and money that saved.

## Fields you need

| Field | Usually in | Spreadsheet column |
| --- | --- | --- |
| The go/no-go call on each bid | RFP platform, a CRM field, or a spreadsheet | Bids: Go/no-go call |
| Who overrode a no-go | Same place as the call | Bids: Overridden by |
| Outcome (won, lost, pending, not submitted) | RFP platform or CRM | Bids: Outcome |
| Outcome date | RFP platform or CRM close date | Bids: Outcome date |
| Deal value, segment | CRM | Deals: Amount, Segment |
| Reason for the call | RFP platform or the go/no-go form | Optional |

## Formula

Count only decided bids (won or lost) with an outcome date in the period.

- **Win rate on backed bids** = bids called go and won ÷ bids called go and decided.
- **Win rate on overridden no-gos** = bids called no-go, overridden and won ÷ bids called no-go, overridden and decided.
- **Calls right** = (go calls that won + overridden no-go calls that lost) ÷ (decided go calls + decided overridden no-go calls). This is the tile value.
- **No-go ledger:** bids called no-go and not pursued, with their value and the reasons. Report the count and value beside the tile, never as part of the accuracy rate.

Record the call before the outcome is known. A call recorded after the result cannot be scored, so leave it out and say how many were left out.

Change against last period: this period's rate − last period's rate, in points.

## Cuts

- All bids, then segment (for example enterprise and public sector).
- Bid type.
- Who overrode the call (sales leader, account executive), for the bid team's own review.
- Reason for the call (fit, capacity, incumbent, price).

## What good looks like

No-gos are seen as strategic focus. The team tracks why it qualifies out and the opportunity cost of each decision, so a no-go reads as capacity saved for the bids it can take.

## The trap

A no-go you cannot quantify looks like lost revenue. So does a team that only ever says no. Show the value and reasons of every no-go, and show the overridden ones that won as well as the ones that lost.

## Not tracked

- **Calls not recorded:** show Not tracked. Start recording the call, its reason and the date on every opportunity before work starts.
- **Overrides not recorded:** show the win rate on backed bids only, and Not tracked for the comparison. Add an "overridden by" field next to the call.
- **Fewer than 10 decided bids in a cut:** show the counts beside the rates, or use the trailing four quarters and say so.

## Outputs

Source: `references/outputs.md`

# Outputs

Two outputs from the same numbers: a one-page Artifact in Claude, or the filled spreadsheet anywhere. Then the weekly refresh.

## One-page Artifact (in Claude)

Publish one page with the Artifact tool, laid out like the deck's dashboard slide ("One page, all eight numbers").

**Header:** "[Company] bid dashboard", the period against last period, the refresh date and who it is for.

**Tiles,** two rows of four, in this order:

| Tile | Value | Unit line | Change |
| --- | --- | --- | --- |
| Pipeline influenced | $ | this period | % against last period |
| Revenue leakage | $ | slipped away | % |
| Go/no-go accuracy | % | of calls right | points |
| SME bottleneck tax | dollar-days | waiting on SMEs | % |
| Content use | multiple, such as 1.5× | win rate with library content | against last period |
| Requirement gaps | $ | lost to named gaps | % |
| Churn from RFPs | % | of renewal ARR | points |
| War room | $ | ACV on the must-win list | % against the last saved run |

Each tile shows its value, the change against last period and one line of note from its reference (for example "Backed [x]% won, overridden no-gos [y]%"). Color the change by whether it is good news: leakage, the SME tax, requirement gaps and churn are better when they fall. A tile that cannot be computed shows **Not tracked** in gray with "Start capturing: [field]".

**War-room table** under the tiles: the must-win deals from [War room](war-room.md), largest ACV first, each with stage, ACV, due date, readiness, main risk, the ask and owner. Flag deals with no activity in 14+ days or a single buyer contact. Under it, the share of forecast pipeline expected to need an RFP. The CRO one-pager's must-win table uses the same list.

**Footer:** the definitions agreed in step 3, the source of each number, the join match rate, small-sample notes, and the line "Computed from [sources] on [date]".

Build notes:

- Keep the data as a JSON block inside the page and render the tiles from it, so the refresh only swaps the data.
- Values are text, never images. The page reads on a phone and in light and dark themes.
- Add a trend line to a tile only when earlier runs give real data points for it. Never draw a trend from a guess.
- Save the Artifact URL in the setup. Every refresh republishes to the same URL.
- Keep the page private. Share it only when the user asks.

## Spreadsheet (any client, or on request)

1. Copy `assets/bid-reporting-dashboard.xlsx` and name the copy `[company]-bid-dashboard-[period].xlsx`.
2. On Deals, Bids, Leaks and Renewals, delete the example rows (every row with its Example column filled).
3. Paste the pulled data under the navy headers, in the column order from [Connections](connections.md). Gray headers are formulas: leave them, and fill them down to the last row of data. On Bids, Value (column E) is a formula, so the export goes into A:D and from F onwards.
4. On Dashboard, set the period start and end, how many must-win deals to list (five by default), the look-ahead days and the as-at date for activity flags. Last period fills itself.
5. Recalculate, check that no cell shows an error, and replace the note of any tile whose data is missing with Not tracked and the field to capture.
6. Hand the file back and say which numbers are Not tracked.

The War room tab lists the must-win deals from the Deals tab.

## Weekly refresh

Ask: "Refresh this every week before the forecast call?" The default is weekly, early on the working day before the call, in the user's time zone.

**In Claude:** create a scheduled task (the scheduled-tasks tool) named "Bid dashboard refresh" with a prompt that carries everything a fresh run needs:

- where the setup and the connection map are saved, and the Artifact URL;
- re-pull the same sources, recompute every number with the saved definitions, and republish the same Artifact;
- save this run's must-win list and its ACV with the setup, so the next run can show the War room change;
- if a source fails, publish with Not tracked on the numbers it feeds and say which source failed, instead of reusing last week's figures;
- send, post, email and write back nothing.

Tell the user when the first run is due and how to pause or change the schedule.

**Elsewhere:** give the user the steps: run the same exports, paste them into the spreadsheet, set the new period, and check for errors.

Sending the page or the file to anyone (email, Slack, Teams, a CRM note) is a separate action. Ask for approval each time.

## Pipeline Influenced

Source: `references/pipeline-influenced.md`

# Pipeline influenced

Tile: **Pipeline influenced**. Value: new pipeline on bid-touched deals this period. Change: % against last period. Note: closed-won on bid-touched deals, and velocity with a bid against without.

## What it tells a CRO

Sales leaders run on pipeline and closed-won. Bids already sit inside both, so this number puts the bid team's work in the units the CRO is measured on. It answers one question: how much of the number did a bid touch, and did those deals close faster?

## Fields you need

| Field | Usually in | Spreadsheet column |
| --- | --- | --- |
| Deal ID, account, amount | CRM opportunity | Deals: Deal ID, Account, Amount |
| Status (open, won, lost) | CRM stage | Deals: Status |
| Created date, close date | CRM | Deals: Created, Close date |
| Size band and segment | CRM (or derived from amount) | Deals: Size band, Segment |
| Link from bid to deal | RFP platform project field, or a CRM checkbox | Bids: Deal ID |
| Bid submitted date | RFP platform | Bids: Submitted |

## Formula

- **Bid-touched deal:** a deal with at least one bid linked to it, by the definition agreed in step 3.
- **Pipeline influenced** = sum of amount on bid-touched deals created in the period, whatever their status now.
- **Closed-won influenced** = sum of amount on bid-touched deals won with a close date in the period.
- **Velocity with a bid** = average days from bid submitted to close, on won bid-touched deals closed in the period.
- **Cycle with and without a bid** = average days from created to close on won deals closed in the period, for bid-touched deals against deals with no bid, by size band. Compare like with like: the same size band, the same period.
- **Share of the number** = pipeline influenced ÷ all pipeline created in the period. Show it only when the CRM total is in the pull.

Change against last period: (this period − last period) ÷ last period.

## Cuts

- Size band (small, mid-market, enterprise): the velocity comparison only means something within a band.
- Segment or region.
- Bid type (RFP, RFI, security questionnaire, DDQ).
- New business against renewal and expansion.

## What good looks like

The bid team uses the same pipeline definition as RevOps or Sales Ops and counts the bookings it works on in dollars. When leaders know the team influences pipeline, the team is part of the strategic conversations, and the forecast call asks for its view.

## The trap

The CRO already has a velocity formula: opportunities × win rate × average deal value ÷ sales cycle length. A bid team velocity number built another way gets challenged on the call. Show the cycle days the CRO's formula uses, and agree the definition with RevOps before the number goes on the page.

## Not tracked

- **No link from bid to deal:** show Not tracked. Start capturing the CRM deal ID on every RFP platform project, or a "bid team involved" checkbox on the opportunity. Until then, a match on account name within a date window works as a stopgap; report its match rate.
- **No bid submitted date:** show pipeline and closed-won, and Not tracked for velocity. Start capturing the submitted date on every bid.
- **No deals without a bid in a band:** show the bid-touched cycle only, and say there is nothing to compare against.

## Requirement Gaps

Source: `references/requirement-gaps.md`

# Requirement gaps

Tile: **Requirement gaps**. Value: deal value lost on bids that named a gap or objection this period. Change: % against last period. Note: lost bids naming a gap, the top gap, and the open pipeline asking for it.

## What it tells a CRO

Bid teams see what the market asks for before product or sales do. Counting the gaps and objections on lost bids, then pricing them against open pipeline that asks for the same thing, turns a loss review into forecast risk and a roadmap input.

## Fields you need

| Field | Usually in | Spreadsheet column |
| --- | --- | --- |
| Gap or objection on each lost bid | Loss reason field, win/loss review, buyer debrief | Bids: Gap or objection |
| Gap or requirement on each open bid | RFP requirements, compliance matrix, the bid team's notes | Bids: Gap or objection |
| Outcome and outcome date | RFP platform or CRM | Bids: Outcome, Outcome date |
| Deal value | CRM | Bids: Value (joined from Deals) |

Use one short, consistent name per gap ("No on-premise hosting", "Pricing too complex"), so counts add up across bids and periods.

## Formula

Tag each gap or objection on lost bids, then count it.

- **Lost bids** per gap = lost bids with an outcome date in the period that named the gap.
- **Lost value** per gap = sum of deal value on those bids.
- **Open pipeline asking** per gap = sum of deal value on open (pending) bids whose requirements ask for the same gap.
- **Tile value** = lost value across all gaps in the period.
- Rank gaps by lost value, and show lost bids and open pipeline beside each.

Change against last period: (this period − last period) ÷ last period.

## Cuts

- Gap (the main cut), ranked by lost value.
- Kind of gap: product, legal, pricing, security, or a sales issue.
- Segment and region.
- New gaps this period against gaps that repeat.

## What good looks like

The team flags gaps as forecast risk on open deals, then rolls them up for product. It sends the gaps out as a short market report each quarter, with the lost value and the open pipeline beside each one.

## The trap

Gaps are more than product. A gap can be legal terms, pricing, a missing certificate, or a sales issue. Tag the kind as well as the gap, or every loss lands on the product team. Staying silent on a gap is how a team loses several bids for the same reason.

## Not tracked

- **No loss reasons on bids:** show Not tracked. Add a gap or objection field to every lost bid and fill it at the win/loss review.
- **Gaps recorded as free text:** group them into named gaps with the user once, save the list in the setup, and reuse it.
- **No requirements data on open bids:** show lost bids and lost value, and Not tracked for open pipeline asking.

## Revenue Leakage

Source: `references/revenue-leakage.md`

# Revenue leakage

Tile: **Revenue leakage**. Value: deal value lost or delayed at a hand-off this period. Change: % against last period. Note: the hand-off that leaked the most, and the split between lost and delayed.

## What it tells a CRO

It shows where deal money slips away between sales, bid, legal and delivery. No sales leader turns down a way to make a deal close sooner, and a leak priced in dollars gets fixed faster than one described as a process problem.

## Fields you need

| Field | Usually in | Spreadsheet column |
| --- | --- | --- |
| Deal ID, account, deal value | CRM | Leaks: Deal ID, Account, Value |
| Lost or delayed | CRM stage history, close-date changes | Leaks: Leak type |
| Days delayed | CRM close-date history, or the bid team's log | Leaks: Days delayed |
| Hand-off reason | Loss reason field, or a tag the bid team adds | Leaks: Hand-off reason |
| Date recorded | CRM or the log | Leaks: Date recorded |
| What happened | Notes, win/loss review | Leaks: What happened |

## Formula

Tag every leak with one hand-off reason, then price it:

1. **Sales to bid:** bad data or a late brief.
2. **Bid to legal:** rework or a slow review.
3. **Negotiation:** days added to the close.
4. **To delivery:** gaps in the hand-over to the delivery or customer success team.

- **Leakage** = sum of deal value on deals lost or delayed at a hand-off, with a date recorded in the period.
- Report **lost** and **delayed** separately as well as together.
- For delayed deals, also report **days added × deal value**, the same unit as the SME bottleneck tax.
- Use the team's own reason list if it agreed one in step 3, and keep it the same every period so the trend holds.

## Cuts

- Hand-off reason (the main cut).
- Segment and size band.
- Lost against delayed.
- Owner, for the bid team's own review only. Leave names off the CRO page.

## What good looks like

The team tracks what it promised in the proposal all the way to renewal. Won deals get a hand-over, and the buyer's requirements reach the customer success team in writing. What a proposal promises shows up later in net revenue retention.

## The trap

A deal won on overpromising becomes future churn. Every overcommitment gets paid back in service credits or a lost renewal, so a leak tagged "to delivery" often surfaces a year later in the churn and contraction tile. Check the two together.

## Not tracked

- **No loss reasons or close-date history:** show Not tracked. Start a leak log with the columns above and fill it in at every win/loss review.
- **Reasons recorded but not by hand-off:** map the existing reasons to the four hand-offs once, with the user, and save the mapping in the setup.
- **Delays not recorded:** show lost value only, and say delays are Not tracked until close-date changes are kept.

## Rfp Sales Report

Source: `references/rfp-sales-report.md`

# RFP sales report for the CRO

Builds the weekly report a proposal leader hands a CRO or VP Sales before the forecast call. It reports proposal work in the sales leader's units (bookings, forecast accuracy, conversion, velocity, efficiency).

Trigger on: "RFP sales report", "CRO report", "proposal team revenue report", "report to sales leadership", "build the proposal scorecard", "what should I show my CRO", "proposal team QBR numbers", or any request to turn CRM plus proposal-tool data into a revenue report for sales leaders.

Do NOT use for bid/no-bid on a single RFP, writing proposal content, or extracting RFP requirements.

## Why the report is shaped this way

- CROs are judged on hitting the number and on calling it early. Forecast credibility protects their tenure, so predictability leads.
- The layout mirrors how revenue leaders run their org: a forecast view (target attainment, coverage, commit, most likely, best case), roughly five must-win deals that represent the quarter, and metrics grouped as predictability, productivity and growth.
- The proposal-specific metrics are revenue generated, no-go rate with capacity saved, shortlist rate, go/no-go accuracy, content use against win rate, and SME bottleneck days, framed around the shift from completion to conversion.

## Step 0: First-run setup (skip if a profile exists)

Look first for a saved "RFP sales report profile" for this organization: a Notion page, a Doc artifact, or `rfp-sales-report-profile.md` in a connected folder. If one exists, read it, use it throughout, and ask only about anything marked unknown or older than two quarters. If none exists, run this setup before building the first report.

### 0a. Ask for documents first

These show what sales leaders push on, which often differs from what they say matters:

- Two or three recent forecast or pipeline call recordings or transcripts.
- The last QBR or board deck, plus its transcript if one exists.
- Any report the proposal team already sends to sales, and how it was received (read, ignored, challenged).
- Win/loss debriefs or evaluator feedback from the last two quarters.

From transcripts, extract into the profile: the questions the sales leader asks repeatedly; which numbers get challenged and which get skipped; the leader's own vocabulary (for example commit, upside, slipped, coverage, Pwin); and every moment proposals or RFPs come up, tagged as risk, blocker or win. Quote short phrases only, attributed by role. If no transcripts are available, say so in the profile and lean harder on the interview.

### 0b. Interview (AskUserQuestion, up to 4 questions per round; plain text if unavailable)

Skip anything the documents already answered. Group the rounds:

1. **The sales leader:** who is the primary reader (CRO, VP Sales, CSO, regional VPs, CEO); what they are paid on this year (bookings, NRR, margin, new logos); what they asked in the last forecast call that nobody could answer; whether they want the number, the risk or the story first; and whether they read a page, skim on a phone, or want it walked through live.
2. **Rhythm and timing:** when the forecast call happens and the quarter calendar; whether this is a weekly pre-read, a monthly review or a QBR section; and whether RevOps already sends a pack this should slot into.
3. **Definitions:** how a deal is marked RFP or proposal-supported in the CRM today; stage names and forecast categories; what counts as a win (signed, shortlisted, awarded); where go/no-go decisions are recorded and whether reasons are captured; whether SME wait time is tracked or only replies; and a rough loaded cost per hour for proposal, SE and legal time.
4. **Systems and data:** which CRM and proposal tool, and whether both are connected; who owns RevOps definitions and will sign off; which fields are reliable and which are filled in carelessly; and any account names or figures that cannot appear in the report.
5. **Politics and goals:** the one thing the team wants leadership to believe about it in six months; where the team gets blamed today (turnaround, losses, cost); which metric would embarrass them if shown honestly; who on the sales side already champions the team and who is skeptical; and a past win or no-go that saved real money, usable as a reference story.
6. **Context checks:** commercial or GovCon (GovCon switches vocabulary to Pwin, recompetes, B&P cost); new logo or renewal heavy; and rough RFP volume and average deal size per quarter, to judge whether segment-level win rates are statistically meaningful.
7. **Product and marketing hand-offs:** who in product receives gap notes and in what format (Jira, Productboard, email, Slack); who in marketing owns messaging and AEO/GEO keywords; the trend window to count over (default trailing four quarters); and whether the company website and content library can be checked for coverage of new requirements.

### 0c. Save the profile

Ask where to save it (Notion page, Doc artifact, Word Doc, SharePoint, or a markdown file in a connected folder), show the content, and save only after the user approves. Sections: readers and what they are paid on; cadence and delivery format; agreed definitions and RevOps sign-off status; systems, reliable fields and exclusions; leader vocabulary and recurring questions from transcripts; team goals, sensitivities and reference stories; context checks; open unknowns. Date the profile.

### 0d. Let the profile shape the report

- Use the leader's own words in headings, tile labels and the headline.
- Order sections by what the leader challenges most; if transcripts show the CRO opens with risk, put must-win deals above the KPI tiles.
- Answer the recurring unanswered question somewhere on the page, or show it as Not tracked with how to start capturing it.
- Drop or move to an appendix any metric the leader never engages with.
- If RFP volume is too low for segment win rates, show trailing four-quarter figures and say why.

## Step 1: Find the data sources

Silently check what is connected (ToolSearch, ListConnectors, listed MCP tools):

- **CRM:** HubSpot, Salesforce, Microsoft Dynamics, Pipedrive, Zoho, or similar. Needed: opportunities/deals with amount, stage, close date, owner, account, forecast category, closed won/lost, loss reason, and any RFP or proposal flag.
- **Proposal tool:** Loopio, Responsive, AutoRFP.ai, Arphie, Inventive, Tribble, 1up, QorusDocs, Ombud, Upland Qvidian, or similar. Needed: projects/RFPs with account, created and submitted dates, status, go/no-go decision and reason, outcome, shortlist/advance status, content used, SME assignments and response times. For trend tracking, also pull requirement-level data: each RFP question or requirement, the answer status (yes, partial, no, roadmap, workaround, custom), the answer text, and any compliance-matrix or gap notes. If the tool doesn't expose questions, ask for the RFP documents or compliance matrices for the period.
- If either is missing, run SearchMcpRegistry for it and offer SuggestConnectors. If none exists, ask for a CSV export of the same fields and carry on.

## Step 2: Settle definitions before pulling data (one AskUserQuestion, up to 4 questions)

Skip any question the profile already answers. Ask only what the systems and profile cannot answer:

1. **Reporting period and cadence:** this quarter to date (weekly view) or last full quarter (QBR view)?
2. **How a deal counts as proposal-supported:** a CRM field or checkbox, a link between proposal project and opportunity, or matching by account name within a date window. Recommend the CRM field if one exists.
3. **Target and commit:** pull from the CRM forecast, or the user supplies the RFP-influenced target?
4. **Where go/no-go decisions live:** proposal tool, CRM stage or field, or a spreadsheet.

If nobody is there to answer, use the recommended option, state the assumption at the top of the report, and continue. Ask the user to confirm the definitions with RevOps; put them in the report footer.

## Step 3: Pull, join and compute in code

Pull records into files in the working directory and compute in Python. Never estimate figures in prose.

Join proposal projects to CRM opportunities using the agreed method. Report the match rate and list unmatched records in a working file.

Compute, for the period and for the prior period (trend):

**The five headline numbers**
1. **Bookings contribution:** sum of closed-won amount on proposal-supported deals; also as % of total bookings in period. (CRO lens: bookings vs plan)
2. **Go/no-go accuracy:** win rate on pursuits marked go, against the share of declined pursuits later won by anyone (if tracked) or the historical win rate of comparable declined deals. (CRO lens: forecast accuracy)
3. **Shortlist rate:** submitted RFPs that advanced to demo, orals or shortlist, divided by submitted RFPs with a known result; plus close rate of shortlisted deals. (CRO lens: stage conversion)
4. **SME bottleneck days:** average business days commit-stage deals waited on SME input (product, legal, security), from proposal-tool assignment timestamps. (CRO lens: deal velocity)
5. **Cost per win and selling hours returned:** pursuit cost (team hours x loaded rate the user gives, or a stated default) divided by wins; hours returned from reuse, automation or self-serve if the tool reports them. (CRO lens: sales efficiency)

**Forecast view:** closed won, commit, most likely, best case for proposal-supported deals against target; pipeline coverage (open pipeline / remaining target); forecast accuracy last period (day-one call vs actual).

**Must-win deals:** the five largest open proposal-supported deals in commit or most likely, each with stage, ACV, due date, readiness (on track / watch / at risk), main risk, and the specific ask. Flag deals with no activity in 14+ days or a single buyer contact.

**Predictability, productivity, growth** (three rows each):
- Predictability: forecast accuracy on proposal deals, win rate on submitted RFPs, deals slipped from commit.
- Productivity: RFP-sourced pipeline created, median proposal cycle time, selling hours returned.
- Growth: win rate new logo vs renewal/expansion, expansion proposals sent, competitive win rate against the most frequent competitor.

**No-go ledger:** count of declined RFPs, team hours not spent, pursuit cost avoided, top no-go reasons, and how many declined deals were later won by anyone.

**Win/loss and market signals (up to four):** most frequent competitor and record against them; content correlated with wins (win rate with vs without a content item, only if sample is 10+); top loss reason; a segment sending unexpected RFPs.

**Decisions needed:** up to three asks for the CRO, each tied to a dollar amount or a must-win deal.

### Trend tracking: product gaps and new requirements

Keep a running trend log next to the profile (`rfp-sales-report-trends`, same location), so counts build across runs. Each entry records the theme, every RFP and opportunity it appeared in, dates, and buyer wording. Never count the same RFP twice for one theme.

Group requirements into themes by meaning, not exact wording ("SOC 2 Type II", "SOC2 report" and "independent security audit" are one theme). Show the grouping in the working files so a person can split or merge themes.

**Product gaps.** A gap is a requirement answered no, partial, roadmap, workaround or custom, or a loss reason or debrief that names a missing capability. A theme becomes a **product trend once it appears in more than two RFPs (3+)** inside the trend window. For each trend, build a product note:
- The gap in one line, plus up to three short buyer phrasings, quoted from the RFPs.
- Count of RFPs, first and last seen dates, and segments or industries affected.
- Logos: every account name tied to the gap (anonymize only if the profile says so).
- Revenue: ACV of affected deals split into lost, open and won-despite-the-gap, from the CRM. Lost and open together are "revenue at stake."
- Competitors in those deals who met the requirement, if known.
- What the proposal team answered (the workaround or roadmap line), so product sees the current story.
Rank trends by revenue at stake.

**New requirements for marketing.** A new requirement is a theme that did not appear in RFPs before the trend window, or that has no approved answer in the content library. It becomes a **marketing trend once more than two RFPs ask for it (3+)**. For each trend, build a marketing note:
- The requirement in one line, the count, first and last seen, and segments.
- The buyers' exact phrasings, written as the questions buyers ask, because those are the queries to target for AEO and GEO.
- Suggested keywords and question-led phrases drawn only from buyer wording, not invented.
- Coverage check: whether the company website and content library already answer it (check the public site with WebFetch when the profile allows). Mark covered, partly covered or missing.
- Win rate on deals that asked for it, if the sample is 10+.
Rank trends by how many RFPs asked and how recently.

A theme below three RFPs stays on a watch list in the trend log and isn't reported as a trend.

Win rate: count submitted proposals only and cut by segment (new logo vs expansion, deal size band). Never show a blended figure alone.

## Step 4: Build the report

Have Artifact list types. Use the Design type (or Data type if the user wants a live dashboard). One fluid page, sections in this order:

1. Header: period, week of quarter, a one-sentence headline stating the forecast position and the biggest risk in dollars.
2. Five headline KPI tiles: value, trend vs prior period, one-line note, and "CRO lens" label.
3. Forecast bar view with coverage and our call.
4. Must-win deals table.
5. Predictability, productivity, growth panels.
6. No-go ledger and win/loss signals side by side.
7. Product gaps and new requirements: two compact panels listing each 3+ trend with its count, revenue at stake (product) or coverage status (marketing), and a pointer to the full note.
8. Decisions needed this week.
9. Footer: definitions agreed with RevOps, data sources, match rate, and framework credit.

Rules:
- Every number comes from computed data. Where a metric cannot be computed, show "Not tracked" with one line on what to start capturing. Never invent a figure.
- Lead with dollars, points of win rate, days and percent of forecast. Volume counts (RFPs submitted, questions answered) go in the appendix if at all.
- Anonymize account names only if the user asks.
- If the user has a brand or design system, use it.

## Step 5: Deliver

Share the artifact with one line on the headline finding, list any metrics shown as Not tracked, the join match rate, and the assumptions made. Deliver the product note and the marketing note as separate, ready-to-send pieces (Docs, or in the format the profile names), one section per trend. Show them to the user and do not send them to product or marketing, or post them to Jira, Slack or email, without explicit approval. Offer a Slides version for a QBR as the next step. Do not post, email or write back to the CRM or proposal tool without explicit approval.

## Sme Bottleneck

Source: `references/sme-bottleneck.md`

# SME bottleneck tax

Tile: **SME bottleneck tax**. Value: deal value × days waiting for an SME, summed this period (in dollar-days). Change: % against last period. Note: average days waiting per bid and the team with the most days.

## What it tells a CRO

Every day a deal waits for an expert is a day it can slip or be lost. Waiting days on their own sound like a scheduling complaint. Waiting days multiplied by deal value are revenue at risk, and that is a number a CRO acts on.

## Fields you need

| Field | Usually in | Spreadsheet column |
| --- | --- | --- |
| Bid, linked deal and value | RFP platform and CRM | Bids: Bid ID, Deal ID, Value |
| Days waiting on product | RFP platform task dates (assigned to answered) | Bids: SME days: product |
| Days waiting on legal | RFP platform tasks, or the legal queue | Bids: SME days: legal |
| Days waiting on security | RFP platform tasks, or the security team's queue | Bids: SME days: security |
| Due date | RFP platform | Bids: Due |

Days waiting = days from an SME task being assigned to it being answered. Count business days if the task dates allow, and say which in the footer.

## Formula

- **Tax for a bid** = deal value × days waiting for an SME on that bid.
- **Tax by team** = sum over bids of deal value × that team's days waiting (product, legal, security, or the team's own list).
- **SME bottleneck tax** = sum of the team taxes on bids due in the period. This is the tile value.
- **Average days waiting** = total SME days ÷ bids with any SME wait, for the note.

Example of the unit: a deal worth 400,000 that waits 6 days on product carries a tax of 2,400,000 dollar-days.

Change against last period: (this period − last period) ÷ last period.

## Cuts

- Team (product, legal, security, others the team names). This is the main cut.
- Commit-stage deals against the rest, since a wait on a commit deal moves the forecast.
- Segment and size band.

## What good looks like

The bid team reports SME delays as dollars at risk, by team, every period. The SME leaders agree the measure before it goes on the page.

## The trap

The tax lands on named teams. Shown without warning, it reads as blame and the SMEs stop helping. Get the SME leaders on board before the first report, and thank SMEs when their days come down.

## Not tracked

- **No SME task dates:** show Not tracked. Start assigning SME questions as tasks in the RFP platform (or wherever SME tasks live) so assigned and answered dates are kept.
- **Tasks kept but not by team:** tag each SME with a team once, in the setup, and reuse the tags.
- **No linked deal value:** show days waiting only, and Not tracked for the tax until bids carry the CRM deal ID.

## War Room

Source: `references/war-room.md`

# War room

Tile: **War room**. Value: combined ACV of the must-win deals. Change: % against the last saved run. Note: how many are at risk, on watch and flagged, and the share of forecast pipeline expected to need an RFP.

The war-room table sits under the tiles: the must-win deals, one row each. The CRO one-pager's must-win table uses the same list.

## What it tells a CRO

The must-win deals are the CRO's forecast for the quarter. Knowing them by name, each with an owner, a readiness call, its main risk and a specific ask, every week, lets sales leadership step in while there is still time. The share of forecast pipeline that will need an RFP tells the bid team how much capacity to plan for.

## Fields you need

| Field | Usually in | Spreadsheet column |
| --- | --- | --- |
| Account, ACV (amount), status | CRM | Deals: Account, Amount, Status |
| Stage | CRM | Deals: Stage |
| Forecast category (commit, most likely, best case) | CRM | Deals: Forecast category |
| Owner | CRM | Deals: Owner |
| Last activity date | CRM activity history | Deals: Last activity |
| Buyer contacts on the deal | CRM contact roles | Deals: Buyer contacts |
| Readiness (on track, watch, at risk), main risk, the ask | The bid team's weekly review | Deals: Readiness, Main risk, The ask |
| Due date of the live bid | RFP platform | Bids: Due (Deals: Next bid due is computed) |
| Expected to need an RFP | CRM field, or the account executive's view | Deals: Needs an RFP |

## Formula

- **Must-win deals** = the five largest open bid-touched deals in commit or most likely, by ACV. Use the forecast categories the CRO uses if they are named differently, and keep the count at five unless the setup says otherwise.
- **War-room table**, one row per must-win deal: deal, stage, ACV, due date, readiness (on track, watch, at risk), main risk, the specific ask, and owner.
- **Flags:** flag a deal with no activity in 14 or more days, or with a single buyer contact.
- **Tile value** = sum of ACV on the must-win deals. The note counts the deals at risk, on watch and flagged.
- **Share of forecast pipeline expected to need an RFP** = value of open deals in commit, most likely or best case marked as needing an RFP ÷ value of all open deals in those categories.

The list is a snapshot, so the change compares with the last saved run. On the first run, leave the change out and save the list with the setup.

## Cuts

- Readiness: at-risk deals first in the note, the table stays sorted by ACV.
- Forecast category: commit before most likely.
- Owner, to spot a person carrying several must-win deals.

## What good looks like

Every must-win deal has an executive sponsor. The bid team brings a readiness call on every commit-stage bid to the forecast call, and the C-level gets a short update on the state of play each week.

## The trap

A war room can turn into a status meeting. Keep it to the ask each deal needs from sales leadership, and make sure sales leadership is in the room to make the call.

## Not tracked

- **No readiness, main risk or ask:** show the deals with stage, ACV and due date, and Not tracked in the empty columns. Start recording all three in the weekly review.
- **No last activity date or contact roles in the CRM:** leave the flags out and say they are Not tracked.
- **No "needs an RFP" field:** show Not tracked for the share. Add a checkbox on the opportunity, or ask account executives to mark commit, most likely and best-case deals each week.
