HG Insights|Phoenix

Skill: SDR Trigger Prioritization

An SDR's 200-account list, ranked to 25 — anchored on the HG triggers that actually pay, not the ones that look exciting in a dashboard.

Overview

Teach Claude to rank an SDR/BDR's outbound triggers the way a competent rep would on a Monday-morning prospecting list. Six trigger families, the recency window that governs each, the dominance order when multiple fire on one account, and the contradiction patterns that look strong on paper but are research-stage churn risk in disguise — so outbound workflows compose from one heuristic instead of re-deriving it inline.

Use cases

  • Monday morning that ends with 25 named accounts, not 200 vibes

    Your SDR opens the prospecting list with 200 named accounts. The skill applies the dominance order — intent in the last 30 days beats a competitor add 11 months old; a high `departmentUsageShare` in the buying function escalates a mid-tier account; a long-tenure incumbent install outranks a soft firmographic match. 25 accounts remain at the top with the dominant trigger named per row.

  • Same heuristic, every SDR/BDR workflow

    Six compound-trigger workflows in the SDR/BDR cohort (anchor prioritizer, incumbent-tenure displacement sequencer, function-led buying signal, SEC-event hook, momentum-decline displacement, and the scorecard) all compose this skill — so the threshold for "strong signal" is identical across every output. No per-workflow drift in the rubric, no re-deriving the recency cutoff inline.

View full skill

SDR Trigger Prioritization

When to use

  • An SDR/BDR workflow is compressing a 200-account prospecting list to a 25-row ranked output.
  • A workflow is asked "which trigger leads on this account" when multiple fire.
  • An author is about to inline trigger-ranking rules into an outbound workflow — stop, reference this skill.

Tools you'll touch

  • search_companies — the cohort-builder primitive. Filter by technology installs, momentum direction (trends[].direction = "declining" | "growing"), firmographic gates. Returns a domain list to fan out per-account probes against.
  • company_intent — per-account top-N topics with buyers_journey (string: "Researching", "Evaluating", "Purchasing") and intent_level. Endpoint is extra: "forbid" — filter post-fetch.
  • list_intent_topics / intent_category — resolve a category to its canonical topic_ids and run the cross-account search.
  • company_install_time_series — per-account adoption intensity over a closed enum window (last_6_months / last_12_months / last_24_months / last_36_months).
  • company_technographic — per-account current-snapshot tech footprint with firstVerifiedDate per row; competitor adds inside the lookback are tech-stack-add triggers.
  • company_technographic.firstVerifiedDate as a renewal-proximity proxy. The contracts tool in HG MCP is scoped to GSI/ICT outsourcing (Accenture, IBM, Cognizant deals), NOT individual SaaS product renewals — the canonical workaround for product-renewal proximity is incumbent install tenure (long-tenure incumbents are in or near their renewal cycle).
  • company_fai — per-account Functional Area Intensity (point-in-time departmentUsageShare by product/department). NOTE: this tool is point-in-time only; there is NO 6-month delta primitive in HG MCP. A hiring shift can be inferred indirectly via a recent technographic add owned by the function, but FAI alone does not expose a trend series.
  • company_firmographic — per-account industry, employee band, revenue band — gating filter, not a primary trigger.
  • sec_filing_section — 8-K (material change) within 90 days; 10-K Item 1A/7A (Risk Factors, MD&A) within 365 days. Requires companyTicker input — workflows querying SEC must supply tickers or resolve them upstream.

The 6 trigger families

#FamilyTool / fieldRecency window
1Buying intentcompany_intent.topics[].score + topics[].buyers_journey≤ 30 days at high resolution; ≤ 90 days at category resolution
2Momentum / adoption trajectorycompany_install_time_series.intensity_momentum and current_intensity deltas; cohort-build via search_companies with trends[].directionwindow-driven (6m / 12m windows)
3Incumbent-tenure renewal proximitycompany_technographic.firstVerifiedDate ≥24 months tenure on the incumbent (the canonical proxy — see "Tools you'll touch")tenure ≥ 24 months
4Tech-stack add (competitor or adjacency)company_technographic.firstVerifiedDate inside lookback; cohort-build via search_companies with technologies≤ 12-24 months
5Function-led buying signal (workaround)company_fai departmentUsageShare for the buying function + company_technographic.firstVerifiedDate on function-owned product adds. NO native hiring-shift delta exists; this composes the two point-in-time signals.function-product add ≤ 12-24 months
6SEC-filing eventsec_filing_section (8-K material change, exec change, M&A, guidance) by companyTicker≤ 90 days (8-K); ≤ 365 days (10-K)

company_firmographic is a gating filter (industry/size band), not a primary trigger. An ICP-matched account with no signal does not belong on the ranked list — it belongs on the SDR's research backlog.

Dominance order when multiple triggers fire

When two or more trigger families fire on one account, lead with the dominant one in the message. The dominance order is:

  1. Buying intent ≤ 30 days at buyers_journey: "Evaluating" or "Purchasing" — the buying motion is active right now.
  2. Incumbent tenure ≥ 24 months on the displacement target — the renewal/decision cycle is open (use company_technographic.firstVerifiedDate since HG MCP does not expose product renewal dates directly).
  3. SEC-filing event ≤ 90 days (8-K) — the account is in transition; the message ties to a public event.
  4. Tech-stack add ≤ 6 months — a recent competitor or adjacency install is leading buying signal proxy.
  5. Momentum decline or rise — adoption trajectory shifts the conversation but rarely leads.
  6. Function-led signal (high departmentUsageShare on a function + recent product add by that function) — modulates which contact to lead with, not which account to lead with.

Two strong signals stacked beat one strong signal alone — but the lead in the opening line is the dominant family, not a paragraph of all five.

Recency cutoffs

Defer to hg-recent-signals for the canonical thresholds. SDR-specific additions:

  • Signal decay window for the opening line is 7-14 days — outside that, the signal is research-context, not the lead.
  • Intent ≤ 30 days at buyers_journey: "Purchasing" is the highest-decay signal — escalate it to "lead today" priority.
  • Tech-stack add ≤ 6 months is the operative window for "recent competitor add"; outside that, frame as "incumbent context", not as a triggering event.

Contradiction patterns (skip the account)

Strong-on-paper trigger combinations that the rubric specifically de-prioritizes:

PatternWhy skip
Rising intent + declining install intensity on incumbentResearch-stage churn risk, not a buying motion — they're evaluating because they're losing the existing system, not because they're funded to add a new one.
Multiple competitor adds in 6 monthsTool sprawl, not consolidation readiness — they're in the middle of an exploration cycle that hasn't converged.
High function-led signal + 10-K layoff disclosure within 90 daysReorg, not growth — net-headcount is the wrong story; lead with the layoff side, not the function-expansion side.
SEC 8-K + intent on categoryConfirm 8-K is on-topic before claiming the link — an 8-K about a CFO change is not a network-security buying signal.

These are not absolute disqualifiers — they downgrade the lead from "open with this" to "research before opening".

Output convention

For an SDR workflow output row:

Account | Dominant trigger (family, recency) | Stacked signals | One-line hook
acme.com | Buying intent, 14d (`buyers_journey: "Purchasing"`) | Tech add (Snowflake, 4mo); Eng `departmentUsageShare` 38% | Snowflake adoption is up + CISO presented at re:Inforce — security tooling RFP is likely live.
  • Dominant trigger column carries family name, evidence, and recency tag.
  • Stacked signals is the supporting evidence (≤2 items).
  • One-line hook follows opening-line-discipline — named/dated event + connection to pain + single targeted question.
  • Every row carries a tool citation per hg-citation-discipline — the source tool name in brackets after each claim.

Common pitfalls

  1. Leading with a stale trigger. A competitor add 13 months ago belongs in the body, not the lead. The 7-14 day decay rule names which signal leads.
  2. Calling buyers_journey: "Researching" "Buying intent". Researching is education-stage — corroborates a buying hypothesis, doesn't anchor it. Per intent-data-calibration, only Evaluating and Purchasing are commercial-trigger-grade.
  3. Inferring an FAI "hiring delta". company_fai is point-in-time — there is no native trend series. The function-led signal must compose departmentUsageShare with a recent product add owned by the function; treating FAI alone as a hiring trigger is fabrication.
  4. Ranking compound signals as the sum of weights. The dominance order is ordinal, not additive — three weak signals do not promote to "Strong" the way one Purchasing-stage intent topic does.
  5. Skipping the contradiction check. Rising intent + declining incumbent usage is the most common false-positive in outbound; document it, don't lead with it.
  6. Passing a domain to sec_filing_section. That tool requires companyTicker. SEC-event workflows must take tickers directly or compose with a ticker-resolution step.

Reference