River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Consulting Data Request List Template

Three documents and two sheets that name the export mode and the expected count for each system before the request goes out.

Free download  ·  No account needed

Format Specification · how to read one row

Two cells decide whether the file that comes back can be trusted

ColumnWhat it has to name
Export Mode To RequestThe export by the source system's own term, Details Only or a saved search, never just "export"
Required FieldsThe columns needed, with the record identifier named even though nobody reading a screen would ask for it
Date Filter FieldThe exact field the range filters on, since most systems carry two or three that are rarely the same population
Expected CountA number the client's own team can already see, read before the export runs
Expected Count SourceWhere that number came from: a list view total, a result count, a summary report
Failure Mode If Asked GenericallyWhat a plain ask actually produces on this specific system

A mode with no count arrives in the right shape and an undercount goes unnoticed. A count with no mode arrives counted and unusable. Both cells or the row is not ready to send.

Search this and every result is an Information Request List: a spreadsheet of what is needed, who owns it, and whether it is open, closed or overdue. The genre's own guides describe it as a status tracker and stop there, new, closed, partially closed, outstanding. None of them say how the export should be produced, so whoever fills the request picks whichever button matches their own screen, and the file that comes back is shaped for reading rather than for joining.

Selkirk Vane Advisory asked Halvergate Marine Supply for the sales pipeline on 3 August 2026. What came back the next day was a Formatted Report grouped by stage: 1,848 rows for 1,842 real opportunities, the extra six being subtotals the grouping inserted. None of the 1,842 carried the eighteen-character record Id anything else in the engagement would need to join against it. The corrected Details Only export, asked for by name on 5 August, did not arrive until 10 August. One system, asked for twice, cost a week.

Jira and NetSuite fail the same way for different reasons. Atlassian's own documentation confirms the classic Data Center export stops at 1,000 issues with no on-screen warning, and its Cloud export article sets a 10,000 batch ceiling there instead. NetSuite's own limit is roughly 20,000 rows, past which a search fails outright rather than returning a partial file. Every request in this pack states the mode and a count taken before the export runs, so a short file gets caught rather than believed.

The two sheets, filled for one engagement across four systems

The Request Register and the Format Specification, filled for Halvergate Marine Supply.

Request Register · Selkirk Vane Advisory for Halvergate Marine Supply

Four systems, one status column that does not lie

ItemSystemData setAgreedStatus
DR-01SalesforceOpportunities, FY202504 AugArrived complete
DR-02Jira Service MgmtSupport cases, FY202524 AugOutstanding
DR-03NetSuiteGL detail, FY202528 AugOutstanding
DR-04SiteLine WMSInventory history, FY202521 AugNo export exists

DR-01's first file arrived complete only after a second ask. The first one, asked generically, came back wrong. That comparison is the next tab.

Format Specification · DR-01, Salesforce · asked twice

Same system, two asks, five days apart

Asked 03 Aug, genericallyAsked 05 Aug, by mode
Export producedFormatted ReportDetails Only, CSV
Rows returned1,8481,842
Real opportunities1,8421,842
Extra rows6 stage subtotalsNone
Record Id presentNoYes, 18-character
Arrived04 Aug10 Aug

DR-02 and DR-03 get the same two cells before they are asked at all: Jira's Data Center export caps at 1,000 issues, so 2,300 tickets go out as three created-date batches; NetSuite's saved search caps near 20,000 rows, so 21,400 GL lines go out as four quarterly searches.

What's in the pack

01

Request Register

Four systems and one status column with six real states, so an incomplete file and a contradicted one read differently from an outstanding one. Each owner is named the way a stakeholder map already names who administers what, not a team. Each access grant listed here is exactly what a kickoff meeting's Milestone Commitment sheet tracks against the date the workplan actually needs it.

02

Format Specification

The export mode, the required fields, the date filter and the expected count, keyed to the same Item ID as the register so neither sheet gets read alone.

03

Export Instructions per System

What Salesforce, Jira Service Management and NetSuite each call their real export modes and batch mechanics, so the general ledger detail NetSuite returns is the same file a financial orientation pass reconciles.

04

Request Letter

One letter organised by system rather than by owner, so a reader forwarding it to their Salesforce administrator can forward just that section.

05

Confidentiality Note

Attached once for everything requested, covering the personal data a payroll extract or a transcript carries without anyone naming it that way first.

06

The Mode-and-Count Rule

The rule behind both sheets: a mode with no count arrives in the right shape and an undercount goes unnoticed, a count with no mode arrives counted and unusable.

07

Verify What Comes Back

Checks a file's shape, its row count against what was expected, and its date filter before the row closes, so a Formatted Report's subtotal rows get named rather than mistaken for a real gap.

08

Chase What Is Outstanding

A graduated ladder, a first chase, a copy to whoever owns the relationship, then a named blocker, firing only on rows genuinely past their agreed date.

How to use it

  1. 1

    Open in River, or take it blank

    Open the pack in River and name the systems the engagement touches, or download the Word documents and CSV sheets and fill them in yourself.

  2. 2

    Inventory before you specify

    Name every system and data set first, the way the intake pass already separates what arrived from what never did, before writing a single export mode.

  3. 3

    Specify the mode and the count

    For each row, name the export by the source system's own term and get a number the client's team can already see, before the request goes out.

  4. 4

    Send, verify, chase

    One letter organised by system, then a shape, count and filter check on whatever arrives, and a chase ladder that only fires once the agreed date has actually passed.

Frequently asked questions

Is this template free?

Yes. The zip is Word documents and CSV sheets, no account and no card needed. Edit with AI is the other branch: it creates a free River account, installs this pack as a private Space, and starts building the register from the systems you name.

What format are the downloaded files?

The three documents come as .docx and the two sheets as .csv, opening in Word, Pages, Google Docs, Excel, Numbers and Sheets with no conversion needed. Add ?format=pdf to the download link for the documents as PDF.

How is this different from an information request list?

An IRL tracks what is asked for and its status: open, closed, overdue. Neither tells the sender how to run the export, which is why a plain ask so often comes back shaped for a screen instead of a file ready to join against everything else already indexed.

Why does Format Specification need both a mode and a count?

Because the two failures are independent. A mode with no count arrives in the right shape and an undercount goes unnoticed inside it. A count with no mode arrives counted against a file shaped for reading rather than for joining. Neither cell substitutes for the other.

What happens once the files actually arrive?

Each one gets checked against its own Format Specification row before Status changes: the fields present, the row count against what was expected, and a spot check against the date filter. The confidentiality terms this pack attaches, covered in a dedicated register for how long each copy is kept, start on the date each file lands, not on the date the engagement ends.

What does 'Edit with AI' actually do?

It reads whatever names the client's systems, from the intake findings or a kickoff conversation, and builds the Request Register from that. Then it works out the export mode and the fields per system, asking rather than guessing wherever a system is not already covered, and drafts the letter once every row is specified.

Ask for the export mode, not just the data

Take the Word documents and CSV sheets blank, or send River the systems this engagement touches and get the register and the format specification back.

Edit with AI