Kit de Outreach — Fase 1

1. Purpose + scope

This SOP operates the 6-stage Evoluciones outreach funnel — from Sales Nav scrape to post-call proposal site. It is the runbook for any work session that touches prospecting, messaging, replies, or qualification.

• Operates: lead acquisition + first conversation + handoff to a discovery call.
• Does not operate: discovery-call delivery itself, contract close, fulfillment of Desarrollo Corporativo™ (EDC) retainers.
• Run by: Daniel (founder, voice owner, qualifier) + agent (research, drafting, ops).
• Consult when: starting a new send batch, drafting messages, reading replies, or onboarding a new agent to the project.

If the SOP and a source document disagree, the SOP wins. Update the SOP in the same commit as any closed decision.

2. Operating principle: 100% them

Every prospect-facing asset is 100% about the prospect — their company, their role, their vertical's specific pain. Evoluciones surfaces only as the answer to a question they're already asking. Daniel's credentials = one trust line in a footer, not a centerpiece.

Applied per stage:

• Cold connect — 100% them, public signals only.
• On-accept DM — ~90% them (their pain) + one credential line + one intervention angle.
• Discovery call — they articulate pain in their words.
• Proposal site — 100% them, with discovery-call data + public signals.

The corporate deck's "Somos un grupo de profesionales..." opening is exactly what hasn't worked for ten years. Do not let it back in.

3. Funnel architecture

1. Sales Nav scrape (Vayne skill)   ← automated via `vayne scrape` subcommand, advanced CSV export
2. LinkedIn connect message         ← public-signals personalization
3. On-accept DM                     ← opens conversation
4. Day-N follow-up + tier-4 link    ← Phase 2 only, if Phase 1 yields signal
5. Discovery call (Cal.com)
6. Custom proposal site (tier-5)    ← post-call, 100% about them

Stage 4 is gated on Phase 1 reply-rate signal. Stages 5–6 are gated on Phase 2 lift. Do not skip ahead — assets built before they're needed waste time and produce generic output.

4. Voice rules

Mirror this voice across every stage — connect, DM, follow-up, proposal:

• Direct second-person (usted or tú per the prospect's apparent register; pick per segment when drafting).
• Quantify pain in dollars, hours, %, or rotation cost wherever possible.
• Specific, not abstract — name systems, behaviors, outcomes.
• Signature framing: "Right now, your X is doing Y. You have no problem generating A; your problem is capitalizing on B."
• Spanish, NLP-coaching idiom (cost of turnover, lost potential, communication breakdowns) — not "revenue ops" tone.
• No "us" preamble. No methodology dump. No service catalog in cold copy.

If a draft reads like the corporate deck, rewrite. If it reads like a create-proposal skill demo (English-speaking ops register), rewrite.

Clustered from ICPRESPONSES.md — Daniel Sr.'s 7 landed conversations + 2 anti-targets. Every pain line below is lifted verbatim from a real conversation (tagged #N), not invented. This is the input to Phase 1 cold copy and Sales Nav targeting.

Source date: 2026-06-04 · Status: Segment A active (Phase 1), B/C written but not targeted.

Strategy locked (why throughput leads)

Of 7 conversations, only #7 closed — and it expanded repeatedly. The other 6 stalled after a positive meeting, all pitched the full Desarrollo Corporativo™ (EDC) retainer. The proven motion:

Land a specific, measurable wedge (the #7 pattern: comunicación efectiva + fijación de metas + coaching individual) → results become visible → client expands to EDC.

Phase 1 targets Segment A (throughput/ops) — the only segment with a close, and the one whose pain converts to a hard $ number. Segment B (banca / client-facing) is closer to revenue but has zero proof → negotiations stall. We build the case studies in A, then attack B in Phase 2 with social proof in hand.

Tone register: usted for first contact (Ecuadorian/LATAM B2B norm). Drop to tú only if the prospect's own register clearly invites it.

Segment A — "Lo que pedí no es lo que hicieron" — LEAD (Phase 1)

Pain archetype: Communication breaks down between people/areas → work comes out different than asked → rework, delays, missed output, cross-area friction. The cost is operational and quantifiable.

Sources: #7 (closed + expanded), #1, partial #2.

Pain in their language (verbatim — use these in copy): - "Las diferentes áreas de la empresa no se entienden y eso está causando problemas de producción, y en el ambiente laboral." — #7 - "Mi equipo no sabe preguntar y el resultado es que hacen algo diferente a lo que el cliente interno esperaba, lo que ocasiona retrasos y problemas con los otros departamentos." — #1 - "Se dan instrucciones incompletas, y después hay reclamos porque no se cumplió como se esperaba." — #2

Buyer / champion: Jefe/Gerente de Talento Humano · Gerente de Operaciones / Gerente de Planta · Jefe de Producción · CIO / Gerente de TI (technical-ops sub-angle, per #1). Economic decision routes upward: HR → gerencia → directorio (#1: "pasó a Talento Humano. y luego a gerencia").

Industries: metalmecánica (#7), alimentos / empacadoras (camarón #1/#2), logística, manufactura en general, equipos técnicos / TI.

Geography: Ecuador, Guayaquil core (all three sources are Guayaquil) → LATAM after.

Company size: ~50–500 (the multi-area handoff problem needs enough departments to exist).

Public signals that flag this pain (LinkedIn / company site): - Manufacturing / logistics / food-processing company, 50–500 employees, multiple departments (producción ↔ calidad ↔ logística ↔ ventas) = handoff surface. - HR/Talent posting about clima laboral, coordinación entre áreas, trabajo en equipo. - Operational scaling signal — new plant/line, hiring surges, expansion announcement → coordination strain. - Both an Ops/Production leader and an HR/Talent role present on the org.

Mapped intervention (the wedge): scoped taller de comunicación efectiva + fijación de metas for one team, + coaching individual for its leads. N sessions, with a before/after rework metric defined up front (so it becomes the case study). NOT the full EDC — that comes on expansion.

Economic / ROI framing (the whole point — see the call framework): personas afectadas × horas perdidas/sem en retrabajo × costo/hora = $Z/mes. The wedge costs a fraction of $Z. This is the RACE "Expand" wedge — #7 proves results drive expansion.

Per-prospect playbook (throughput-specific, layers on SOP §7): from the row, confirm (a) multi-department org, (b) ops/production or HR/Talent seniority, (c) any scaling/coordination signal. Pain hypothesis = "instrucciones que se entienden distinto entre [área X] y [área Y] → retrabajo y retrasos." Connect line names the handoff friction, no offer.

Segment B — "Mis líderes no saben liderar" — PHASE 2 (do not target yet)

Pain archetype: Leadership/management capability gap → turnover, disengaged teams, mechanical service, weak client experience.

Sources: #2, #4, #5, #6. Zero closes — all stalled after positive meetings.

Pain in their language (verbatim): - "Las personas en posiciones de liderazgo necesitan aprender a manejar mejor a sus equipos, para bajar la rotación de personal." — #2 - "La mayoría ha pasado mucho tiempo haciendo lo mismo, y ya lo hacen mecánicamente, lo que los clientes perciben… necesito despertarlos." — #4 - "El equipo de liderazgo (gerentes de hoteles) necesita aprender a motivar al equipo administrativo, y tener las habilidades para poder hacer coaching." — #5 - "La cúpula de la empresa… necesita adquirir habilidades de comunicación, y aprender a delegar." — #6

Buyer / champion: HR / People & Culture / Talent Development / D&I; newly-appointed jefes de área (#4).

Industries: banca preferente/privada (#4), hotelería (#5), servicios, logística (exec layer #6).

Why Phase 2, not now: highest revenue proximity (client retention/experience = RACE Expand), biggest deal ceiling — but no results to back claims, so the buyer has nothing to defend internally and "pauses for other priorities." Re-open once Segment A produces case studies. Note: #4 (banca preferente, 150 staff serving high-net-worth clients) is the single highest-value target for Phase 2.

Segment C — "El equipo se quiebra bajo presión" — EMERGING (n=1, do not target)

Pain archetype: Emotional regulation for high-stress, frontline / customer-exposed roles.

Source: #3 only.

Pain in their language (verbatim): - "Necesitamos que el equipo de brigadistas sepa manejar situaciones difíciles… 'primeros auxilios emocionales', porque les toca tratar con personas que han sido víctimas de asaltos o amenazas." — #3

Mapped intervention: talleres de manejo de emociones + comunicación efectiva + coaching.

Status: single data point — not a segment yet. But #3 is banca/Ecuador and showed high interest ("hay mucho interés"); fold into the Phase 2 banking push as a specialized angle, don't lead with it.

Anti-targets — EXCLUDE from Sales Nav

Both anti-targets are competitors, not buyers:

Exclude these titles/keywords from every search: coach, coaching, NLP, PNL, facilitador, facilitadora, consultor de desarrollo, capacitador, formador, desarrollo humano (as a personal title), bienestar consultancies. Anyone who sells this is competition. Target the buyers (HR/Talent, Ops) — not the practitioners.

Known blind spot: the anti-target slots captured competitors, not skeptical buyers. We have no data on where the framing fails with a real prospect. Harvest that from Phase 1 reply data (SOP §9) and feed it back here.

Buyer-role map (applies across segments)

• Champion = HR/Talent or a department head who owns a team's output (Ops/Production for Segment A).
• Economic decision routes upward: champion → gerencia → directorio (observed in #1, #5).
• Implication for the call: arm the champion to sell upward — the one-page ROI calc from the discovery-call framework exists precisely so they can carry a number, not generalities, into the gerencia/directorio meeting. This is the antidote to the "lo pongo en pausa" stall.

Phase 1 prospecting search for the throughput beachhead. v2 (locked 2026-06-04) — title-based, 3-city, broad-recall. Round-trip-validated live: 693 results.

Search type: Lead (people) · Result-cap risk: n/a (693, comfortably under the 2,500 cap).

Strategy — v2 inverts v1 (precision → recall)

v1 optimized for precision (tight FUNCTION+geo net) to avoid wasting enrichment credits on noise — and bled ~60% off-filter attrition. v2 optimizes for recall because the ICP-fit gate is now a free subagent pass ($0 API), so noise is dropped at no cost and LeadMagic only ever enriches confirmed keepers. A wide, slightly-noisy net beats a narrow clean one: bigger clean-ICP corpus for the same cheap, preview-gated Vayne spend.

Key change: CURRENT_TITLE clusters replace FUNCTION (titles are precise + language-normalized; FUNCTION mis-tagged production heads as "Manufacturing" and leaked coordinators/IT-support). Plus +Manufacturing industry, +headcount F, +Owner/Partner seniority, Ecuador's 3 industrial cities.

The URL (paste into Sales Navigator / feed to Vayne)

https://www.linkedin.com/sales/search/people?query=(filters%3AList((type%3AINDUSTRY%2Cvalues%3AList((id%3A23%2Ctext%3AFood%2520and%2520Beverage%2520Manufacturing%2CselectionType%3AINCLUDED)%2C(id%3A66%2Ctext%3AFisheries%2CselectionType%3AINCLUDED)%2C(id%3A840%2Ctext%3AFabricated%2520Metal%2520Products%2CselectionType%3AINCLUDED)%2C(id%3A116%2Ctext%3ATransportation%252C%2520Logistics%252C%2520Supply%2520Chain%2520and%2520Storage%2CselectionType%3AINCLUDED)%2C(id%3A25%2Ctext%3AManufacturing%2CselectionType%3AINCLUDED)))%2C(type%3ASENIORITY_LEVEL%2Cvalues%3AList((id%3A210%2Ctext%3AExperienced%2520Manager%2CselectionType%3AINCLUDED)%2C(id%3A220%2Ctext%3ADirector%2CselectionType%3AINCLUDED)%2C(id%3A300%2Ctext%3AVice%2520President%2CselectionType%3AINCLUDED)%2C(id%3A310%2Ctext%3ACXO%2CselectionType%3AINCLUDED)%2C(id%3A320%2Ctext%3AOwner%2520%252F%2520Partner%2CselectionType%3AINCLUDED)))%2C(type%3ACOMPANY_HEADCOUNT%2Cvalues%3AList((id%3AC%2Ctext%3A11-50%2CselectionType%3AINCLUDED)%2C(id%3AD%2Ctext%3A51-200%2CselectionType%3AINCLUDED)%2C(id%3AE%2Ctext%3A201-500%2CselectionType%3AINCLUDED)%2C(id%3AF%2Ctext%3A501-1000%2CselectionType%3AINCLUDED)))%2C(type%3ACURRENT_TITLE%2Cvalues%3AList((text%3AGerente%2520de%2520Planta%2CselectionType%3AINCLUDED)%2C(text%3AJefe%2520de%2520Planta%2CselectionType%3AINCLUDED)%2C(text%3AJefe%2520de%2520Producci%25C3%25B3n%2CselectionType%3AINCLUDED)%2C(text%3AGerente%2520de%2520Talento%2520Humano%2CselectionType%3AINCLUDED)%2C(text%3AJefe%2520de%2520Talento%2520Humano%2CselectionType%3AINCLUDED)%2C(text%3AGerente%2520de%2520Recursos%2520Humanos%2CselectionType%3AINCLUDED)%2C(text%3AJefe%2520de%2520Recursos%2520Humanos%2CselectionType%3AINCLUDED)%2C(id%3A203%2Ctext%3AChief%2520Information%2520Officer%2CselectionType%3AINCLUDED)%2C(id%3A81%2Ctext%3ADirector%2520of%2520Operations%2CselectionType%3AINCLUDED)%2C(id%3A584%2Ctext%3AHead%2520of%2520Operations%2CselectionType%3AINCLUDED)%2C(id%3A28%2Ctext%3AOperations%2520Manager%2CselectionType%3AINCLUDED)%2C(id%3A173%2Ctext%3AVice%2520President%2520Operations%2CselectionType%3AINCLUDED)%2C(id%3A108%2Ctext%3AProduction%2520Manager%2CselectionType%3AINCLUDED)%2C(id%3A105%2Ctext%3AHuman%2520Resources%2520Director%2CselectionType%3AINCLUDED)%2C(id%3A45%2Ctext%3AHuman%2520Resources%2520Manager%2CselectionType%3AINCLUDED)%2C(id%3A153%2Ctext%3AChief%2520Technology%2520Officer%2CselectionType%3AINCLUDED)%2C(id%3A163%2Ctext%3ADirector%2520of%2520Information%2520Technology%2CselectionType%3AINCLUDED)%2C(text%3AGerente%2520de%2520Operaciones%2CselectionType%3AINCLUDED)%2C(text%3AJefe%2520de%2520Operaciones%2CselectionType%3AINCLUDED)%2C(text%3ADirector%2520de%2520Operaciones%2CselectionType%3AINCLUDED)%2C(text%3AGerente%2520de%2520Producci%25C3%25B3n%2CselectionType%3AINCLUDED)%2C(id%3A303%2Ctext%3APlant%2520Manager%2CselectionType%3AINCLUDED)%2C(id%3A488%2Ctext%3ASupply%2520Chain%2520Manager%2CselectionType%3AINCLUDED)%2C(id%3A296%2Ctext%3ALogistics%2520Manager%2CselectionType%3AINCLUDED)%2C(id%3A779%2Ctext%3AWarehouse%2520Manager%2CselectionType%3AINCLUDED)%2C(id%3A17%2Ctext%3AGeneral%2520Manager%2CselectionType%3AINCLUDED)%2C(text%3AGerente%2520General%2CselectionType%3AINCLUDED)))%2C(type%3AREGION%2Cvalues%3AList((id%3A117009082%2Ctext%3AGuayaquil%252C%2520Guayas%252C%2520Ecuador%2CselectionType%3AINCLUDED)%2C(id%3A104011438%2Ctext%3AQuito%252C%2520Pichincha%252C%2520Ecuador%2CselectionType%3AINCLUDED)%2C(id%3A104260525%2Ctext%3ACuenca%252C%2520Azuay%252C%2520Ecuador%2CselectionType%3AINCLUDED)))))

Session params (recentSearchParam, sessionId, viewAllFilters) stripped — they're per-session cruft; Vayne uses its own authenticated session.

Filters applied (authoritative spec)

Keywords: none. Spanish free-text titles + English taxonomy IDs both used — LinkedIn ORs them and normalizes localized titles.

Competitor exclusion — moved to the free fit-gate

Title-EXCLUDED clusters are an Advanced-Plus-tier feature (unavailable here). Not needed: the industry filter structurally excludes coaches/PNL trainers, and the ICP-fit subagent drops any competitor title (coach/PNL/facilitador/...) that slips through — at $0 cost.

Dedup — by LinkedIn URN across orders

The corpus must be DISTINCT by member_linkedin_id across the new order + the existing 22-lead order 69d7e647… (already enriched — keep them, don't re-enrich). "Exclude people already contacted" is a send-time filter for the next pass (mark do_not_contact), not a corpus filter — prior-contacted voices still inform the synthesis.

Levers (if the scrape feels too big/small)

• Trim 693 → ~300–400: drop headcount F, or drop INDUSTRY 25 (keep the 4 specific buckets). Or scrape a capped subset via Vayne limit=N.
• Grow: re-add Ecuador-wide REGION (106373116) — but that 1.5k'd it; small-town noise dilutes density.

Operating notes (per SOP §8)

• One scrape per segment per week max. Save raw CSV to references/ (gitignored), ingest via python -m evoluciones_outreach ingest, enrich with ENRICH_EMAIL=false (keepers only).

Segmento: Throughput / Operaciones ("Lo que pedí no es lo que hicieron")

Esto es un marco, no un guion. No lo lea palabra por palabra. Úselo para no salirse del único objetivo de la llamada: convertir el dolor en un número, y cerrar un siguiente paso con fecha.

El error que mató 6 de 7 conversaciones: hablar de "comunicación efectiva" y "desarrollo" en abstracto. El cliente quedó interesado pero sin un número que defender ante su gerencia → "lo pongo en pausa". Esta llamada existe para no repetir eso.

⚠️ Guardia anti-divagación (léala antes de empezar)

Si se escucha hablando de "desarrollo del ser", "PNL", "comunicación efectiva" o cualquier generalidad por más de 30 segundos sin un número, pare y regrese a estas tres preguntas: ¿Cuántas personas? · ¿Cuántas horas se pierden? · ¿Cuánto cuesta eso?

No venda PNL. No venda el EDC completo. Hoy solo: cuantifica → propone una cuña medible → agenda el siguiente paso.

Antes de la llamada (30 seg)

• [ ] Tenga lista su tarifa de la cuña ($____) — sin un precio no puede cerrar el cálculo.
• [ ] Anote la señal pública que usó en el mensaje (qué áreas, qué industria).
• [ ] Abra esta hoja y la calculadora de abajo en blanco.

1. Apertura (1 min) — anclar al dolor, pedir permiso

• Agradezca y recuerde por qué conectaron: "Le escribí porque vi que en [empresa] hay varias áreas trabajando juntas y, en operaciones así, lo que se entiende distinto entre áreas suele costar caro."
• Pida permiso para preguntar: "¿Le parece si le hago unas preguntas rápidas para ver si esto les pasa y cuánto pesa? Si no aplica, se lo digo de una."

Punto de reseteo (téngalo a la vista toda la llamada): si en cualquier momento lleva 30 seg hablando sin un número, vuelva a estas tres → ¿cuántas personas? · ¿cuántas veces a la semana? · ¿cuánto tiempo se va en arreglarlo?

2. Diagnóstico cuantificado (6–7 min) — EL CORAZÓN

Haga estas preguntas en orden. No avance hasta tener un número (aunque sea estimado) en cada [ ]. Si dice "no sé", use el aproximador: "Si tuviera que adivinar, ¿más cerca de X o de Y?"

3. El cálculo EN VIVO (2 min) — dígalo en voz alta

Llene esto mientras habla con él, no después. Que lo vea y lo escuche.

   A  personas afectadas .............. ______
×  F  veces/semana que se rehace ...... ______
×  H  horas por incidente ............. ______
=  ______ horas perdidas / semana
×  4.3 semanas .........................   (≈ mes)
=  ______ horas perdidas / mes
×  $___ costo por hora del equipo .....   (ver aproximador ↓)
─────────────────────────────────────
=  $______  AL MES  ← este es el número (Z)

Aproximador del costo/hora (si no lo sabe, no se quede sin número): - "¿Cuánto gana en promedio al mes una de esas personas?" → sueldo ÷ 160 ≈ costo/hora. - Referencia rápida (Ecuador, ajuste según el caso): operativo $600–1.000/mes → ~$4–6/h · técnico/supervisor $1.500–2.500/mes → ~$9–16/h · profesional/jefatura $3.000+/mes → ~$19+/h. - Si hay varias personas con sueldos distintos, use un promedio. Redondee hacia abajo: un número conservador es más creíble. - Si no sabe ningún sueldo, use la banda por tipo de cargo de arriba. En empresas chicas (bajo 100 personas) tienda al piso de cada banda; en grandes, al techo. Nunca se quede sin número — un estimado conservador basta para seguir.

Dígalo así (ejemplo): "Entonces, tirando por lo bajo, esto les está costando alrededor de $Z al mes — unos $[Z×12] al año — solo en retrabajo, sin contar [lo del punto 5]."

→ Pause. Deje que el número aterrice. No lo tape hablando.

¿Y si el número sale bajo? Si $Z es menor o igual a su precio de cuña, no fuerce la venta — pierde credibilidad y es justo lo que un buen comprador huele. Sea directo: "Por ahora esto no les cuesta lo suficiente como para justificar meterle plata. Cuando llegue a [2–3 veces por semana / X horas], me llama y lo resolvemos." Si en el camino aparece un dolor más caro (rotación, reclamos de clientes, entregas perdidas), cambie a ese y vuelva a calcular. Pero no invente un número para cerrar hoy.

4. La propuesta-cuña (2 min) — chico, medible, no el EDC completo

No proponga el Programa de Desarrollo Corporativo completo. Proponga una cuña que se pueda medir y que se pague sola. (Así cerró y se expandió el cliente de metalmecánica.)

• La cuña: un programa corto y enfocado para ese equipo — comunicación efectiva + fijación de metas, con coaching individual para los líderes del proceso. ____ sesiones (su formato).
• Encuádrelo contra el número: "La inversión es $[su precio] — una fracción de los $Z al mes que ya están perdiendo. Si bajamos el retrabajo aunque sea a la mitad, se paga sola en [____] semanas."
• Defina la métrica ANTES de empezar (esto crea el caso de éxito): "Medimos [veces/sem de retrabajo entre esas áreas] hoy, y lo volvemos a medir en [N] semanas. Si baja, seguimos; si no, no le cobro la expansión."

5. Cierre con FECHA (1 min) — mate el "lo pongo en pausa"

6 de 7 conversaciones murieron en "déjame verlo". No cierre sin un compromiso de calendario.

• "Le armo esto en una sola página — el cálculo que acabamos de hacer, la cuña, y la métrica — para que usted lo pueda presentar internamente. ¿Le llega bien si se la mando [mañana / fecha]?"
• "Y para no perder el hilo, ¿agendamos 15 min el [día/fecha] para revisarla juntos?" → AGENDE AHÍ MISMO.
• Si menciona a gerencia/TH: "¿Quién más necesita verlo, y cuándo se reúnen?" → ate la página de una hoja a esa reunión.

6. Battlecard de objeciones (si aparecen)

Recordatorio final

La PNL y "el desarrollo" son el cómo, no el qué. El cliente compra menos retrabajo, menos retrasos, menos reclamos — un número que pueda defender ante su jefe. Hable de ese número. El resto es relleno.