Skip to content

dimagi-internal/ace / skills/build-targeting-model

build-targeting-model

Verified

Turn a targeting selection into a shareable Doc and a formula-driven Excel model. Use when the answer must leave the conversation.

Build Targeting Model

A figure that only exists in a chat transcript cannot be argued with, and cannot be re-run when somebody challenges the unit cost. This skill produces the two artifacts a funding conversation actually needs: a document that states the claim with its sources, and a spreadsheet that is a model rather than an export — assumptions in input cells, every cost column a formula, so a reader can change a number and watch it propagate.

The distinction matters more than it sounds. A static export invites the reader to trust the total. A model invites them to test it, and the ones that survive testing are the ones that get funded.

Inputs

Every row except the optional Phase 1 PDD is an explicit argument or a live labs read; no run is needed.

SourceArtifactUsed for
Upstreamthe selection and its totalsthe headline figures
defend-a-figurethe three paragraphs and Sources tablethe document body
Connect Labs MCPindicators_select (high limit)the ranked geography rows
Connect Labs MCPcea_market_sizecost basis and unit price
Connect Labs MCPindicators_methodologysource URLs, licences, cross-check verdicts
Phase 1 (optional)1-design/idea-to-pdd.mdthe intervention and its pricing basis
cost-effectiveness (optional)cea_model_run per scenarioa Cost-effectiveness sheet: the ranking as values, with price and units of value per death as yellow input cells and the multiple as a formula (deaths averted x units per death / spend / benchmark), and the uncounted benefits, so the funder can change the price live

Products

  • A Google Doc via drive_create_doc_from_markdown — claim, defence, methods considered, Sources table with live links
  • <country>-<indicator>-targeting-model.xlsx via drive_upload_binary — five sheets, described below
  • <phase>/build-targeting-model_summary.md — what was produced and the links

Terminology

Every string this skill puts in front of a reader follows skills/_terminology.md: Connect, never CommCare Connect, and US spelling, always — program, enrollment, enroll, organization, behavior, center; never programme, enrolment, organisation (owner, 2026-10-20: "everything should be US spelling", ace#3232). Identifiers (case properties, slugs, ids, config keys) keep their spelling.

Process

  1. Confirm the figures have been defended. If defend-a-figure has not run, run it. Publishing an undefended number is the failure this whole toolchain exists to prevent.
  2. Pull the ranked rows with indicators_select at a high limit — the document shows a headline, the model shows the list, and a reader who cannot see their own district loses interest in both.
  3. Settle the cost basis explicitly with cea_market_size. A unit price is meaningless without a unit: per newborn, per child, per household, per case. Where an indicator implies no case count the per-case basis is refused rather than approximated — pick another basis, do not work around it.
  4. Build the workbook with openpyxl (available in the labs venv), as a model and not an export — the sheet layout is below.
  5. Make every cost column a formula referencing the Assumptions cells. A reader who changes the unit price and sees nothing move stops believing the spreadsheet.
  6. Reconcile the column sums against the published totals and disclose the difference. Row values arrive rounded, so a column can differ from the headline by a few units. State it on the Assumptions sheet; a reader who finds it themselves assumes worse.
  7. Write the Doc from the defence, with the Sources table carrying a working URL per input, taken from the methodology rather than invented.
  8. Upload both and return the links together. A model without its document loses its caveats; a document without its model cannot be tested.

MCP Tools Used

connect_labs (remote HTTP, PAT auth)

  • indicators_select — built
  • cea_market_size — built
  • indicators_methodology — built

ace-gdrive

  • drive_create_doc_from_markdown — built
  • drive_upload_binary — built

The workbook

SheetHolds
AssumptionsUnit cost, pricing basis, threshold, method, vintage — highlighted as inputs. Everything else references these. Carries the rounding note.
Ranked geographiesThe eligible list ranked by burden, with population, the count being funded, cumulative share, and formula-driven cost columns.
Threshold comparisonThe same selection at each threshold under discussion, so the reader can see what the cutoff is buying.
Method comparisonWhat every method reports, with the conservative figure named rather than left to be inferred.
Sources & caveatsSource, vintage, licence, live URL, and each cross-check's verdict.

The cost columns are the load-bearing part: =IF(Assumptions!$B$6="birth", E2, ...) to pick the denominator from the basis cell, and =H2*Assumptions!$B$5 for the cost itself. Hard-coded costs are the single most common reason a model gets returned.

Mode Behavior

  • Auto: build both artifacts, upload, notify the admin group with the links and the headline figure.
  • Review: present the document draft and the model's assumptions for approval before uploading to a shared Drive location.

Change Log

DateChangeAuthor
2026-08-29Initial version — model-not-export principle, five-sheet layout, formula-driven costs, rounding disclosureACE team
canvas-designanthropics/skillsCreate beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
docxanthropics/skillsUse this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files) or Word templates (.dotx files). Triggers include: any mention of 'Word doc', 'word document', '.docx', '.dotx', or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx or .dotx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'template', or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
pdfanthropics/skillsUse this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.Stars: 196kUpdated Claude CodeCodexVerifiedInstall
pptxanthropics/skillsUse this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx or .potx file (even if the extracted content will be used elsewhere, like in an email or summary); editing, modifying, or updating existing presentations; combining or splitting slide files; working with templates (.potx), layouts, speaker notes, or comments. Trigger whenever the user mentions "deck," "slides," "presentation," or references a .pptx or .potx filename, regardless of what they plan to do with the content afterward. If a .pptx or .potx file needs to be opened, created, or touched, use this skill.Stars: 196kUpdated Claude CodeCodexVerifiedInstall