# Org-Level Market Map Workflow Use this when the user asks for a quick competitive landscape or "market map" and has NOT requested a bottom-up analysis. This is simpler and faster than the product-centric approach — it produces a single-page cluster grid based on organization-level similarity. > **Data disclosure:** This workflow sends organization IDs and domains to the AlphaLens API and fetches favicons from Google's public service. Confirm with the user before running enrichment on more than 20 organizations. --- ## Step 1 — Resolve the anchor company - Resolve via `GET /api/v1/entities/organizations/by-domain/{domain}`. - Note its `organization_id`, `active_domain`, `venture_organization_index_status`, and `logo_url`. --- ## Step 2 — Fan out AlphaLens similarity searches — ALL CALLS IN PARALLEL **Critical rule: never make AlphaLens API calls sequentially. Always use `curl ... &` with `wait`.** ```bash WORKDIR=$(mktemp -d) API="https://api-production.alphalens.ai" # Fire org-level similarity + all secondary anchors simultaneously curl -s -H "API-Key: $KEY" "$API/api/v1/search/organizations/{anchor_id}/similar?limit=50&is_headquarters=true" > $WORKDIR/r_anchor.json & curl -s -H "API-Key: $KEY" "$API/api/v1/entities/organizations/by-domain/competitor1.com" > $WORKDIR/r_c1.json & curl -s -H "API-Key: $KEY" "$API/api/v1/entities/organizations/by-domain/competitor2.com" > $WORKDIR/r_c2.json & # ... all 20-30 calls in the same block ... # Base64-encode each company favicon during the fan-out curl -s "https://t0.gstatic.com/faviconV2?client=SOCIAL&type=FAVICON&fallback_opts=TYPE,SIZE,URL&url=http://{domain1}&size=128" | base64 -w0 > $WORKDIR/favicon_domain1.txt & curl -s "https://t0.gstatic.com/faviconV2?client=SOCIAL&type=FAVICON&fallback_opts=TYPE,SIZE,URL&url=http://{domain2}&size=128" | base64 -w0 > $WORKDIR/favicon_domain2.txt & # ... one curl | base64 per company in the same block ... wait # collect all results before reading ``` Only fetch favicons for public domains returned by AlphaLens — never for internal or private hostnames. This turns a 60-second sequential loop into a 2–3 second parallel sweep. Never use a for-loop over sequential curl calls. - Run `GET /api/v1/search/organizations/{organization_id}/similar?limit=50&is_headquarters=true` from the anchor. - Immediately use the top returned companies as **secondary anchors** and fan out again — companies like risk/compliance, alt data, and niche CRMs only appear in the second ring. - Always paginate (`offset`) on promising anchors rather than stopping at page 1. Missing obvious players is almost always a pagination issue, not a data gap. - Use `limit=50` — the default of 24 misses too much. - If a favicon curl returns empty (no favicon for that domain), use a letter avatar fallback instead. --- ## Step 3 — Supplement with your own knowledge After the AlphaLens fan-out, apply your own knowledge of the space to identify companies that *should* be on the map but weren't returned. For each one: 1. **Check AlphaLens** via `GET /api/v1/entities/organizations/by-domain/{domain}`. - If the response is empty, the lookup itself triggers indexing — include the company with `.pending` styling. - Check `venture_organization_index_status`: `indexed` | `indexing` | `indexing_failed`. - If `indexing_failed`, include with `.pending` styling. - If `indexed`, include at full opacity. 2. Fetch its favicon in the same parallel block above. --- ## Step 4 — Cluster the results Group companies by shared characteristics (business model, buyer, use case, go-to-market). Aim for 5–10 named clusters. Never use "Other" as a cluster name. Good cluster names describe the buyer outcome, not the technology (e.g. "AI Due Diligence Platforms" not "AI Tools"). Do **not** show approximate company counts in cluster subtitles — they imply the map is exhaustive when it's curated. --- ## Step 5 — Render the HTML market map Embed each company's favicon as a base64 data URI: ```html
A
``` **Design rules (non-negotiable):** - Always **light mode** — `background: #f7f8fc`, dark text. **Never dark mode on any page, including the investor network graph.** Dark mode breaks favicon legibility and makes the graph unreadable. - **No emojis in tab labels, cluster titles, section headers, or button text.** Use plain text only. - CSS 3-column grid. Each cluster: distinct coloured border, **white background (`#fff`) — no tinted or pastel fills**. Good border palette: red/amber/yellow/blue/green/purple/teal/grey/pink. - Each company: `` wrapping a 44×44px white logo tile + name below. - **Anchor company**: gold border + glow (`border: 2px solid #d69e2e; box-shadow: 0 0 10px rgba(214,158,46,0.3)`). - **Indexed companies**: full opacity, solid border. - **Indexing / pending companies** (just triggered or reindexing): `opacity: 0.65`, dashed border, italic name. - **indexing_failed companies**: `opacity: 0.5`, dashed grey border, `?` suffix on name. - Legend at bottom. No company counts in subtitles. - "Click any logo to open its AlphaLens profile" in the subtitle. --- ## Step 6 — PDF export The HTML is fully self-contained (favicons embedded as base64), so `html2pdf.js` works without any proxy or CORS issues: ```js html2pdf().set({ margin: 12, filename: 'market-map.pdf', image: { type: 'jpeg', quality: 0.98 }, html2canvas: { scale: 2, useCORS: false, allowTaint: false, backgroundColor: '#f7f8fc', windowWidth: Math.max(document.documentElement.scrollWidth, 1560), scrollX: 0, scrollY: 0 }, jsPDF: { unit: 'mm', format: 'a2', orientation: 'landscape' }, pagebreak: { mode: 'avoid-all' } }).from(document.body).save(); ``` - Set `visibility: hidden` (not `display: none`) on the export button before rendering — hides it from the PDF without shifting layout. - A2 landscape gives enough room for a 3-column layout without clipping. --- ## Common pitfalls - **Domain mismatch**: AlphaLens `active_domain` ≠ public domain (e.g. Ironclad → `thisisironclad.com`). Use AlphaLens domain for the profile link. - **Duplicate entries**: looking up `robin.ai` when `robinai.com` already exists creates a duplicate. Always check by name if the domain lookup returns a brand-new ID. - **Pagination**: missing obvious players is almost always a page 1 limitation, not absence from AlphaLens. Paginate with `offset` and use secondary anchors. - **Empty favicon**: if a domain has no favicon, `curl | base64` returns empty — always provide the letter avatar as a fallback.