# B_01 — Macro Strategy: Research Doc vs Execution Reality **Sub-agent**: B1 of 8. Cross-references `cloud-partner-niche-research.md` (2026-04-24 strategy doc) against Avishay-stated direction (`04_avishay_synthesis.md`), actual decay (`02_decay_verdict_by_campaign.md`), and Instantly + leads-DB ops. **Tier legend**: A = direct evidence + ≥500 sends or unambiguous primary doc. B = one strong signal + corroborating context. C = inferential, single source. --- ## Strategy-vs-Execution Matrix | # | Strategy-doc recommendation | Verdict | Evidence | Tier | |---|---|---|---|---| | 1 | "Wedge niche 60/40 cloud-partner" (research-doc line 11, 1735) | **VALIDATED** | Actual cloud-partner share of lifetime sends = **61.9%** (142,916 / 231,000); replies 71.2%; opps 79.7%. Almost exactly the 60/40 target. Source: `instantly_lifetime_stats.jsonl`. | **A** | | 2 | "Sweet-spot ICP: 20-150 emp IL+EU partners, Premier-chasing, no SDR" (research-doc lines 67, 890-893) | **DISCONFIRMED in execution** | ICP definition assumes partners are *GG's customers*. Real ops: EC/WideOps IS GG's only cloud-partner client; the "Premier Partner Credits" + "Cloud Partners" campaign LISTS are actually startups (Kaapa Biotech, Voltrac, Saga, Adaptyv — `leads_snapshot.db` `GG - Premier Partner Credits - EU DM - 2026-04` n_companies=266). Campaigns "MSPs Acquisiton Test", "Cloud Reseller - Growthgrid", "IL Cloud Partners — 2026-04-19 (HeyReach)", "Growthgrid - US Cloud Partners - AWS", "Growthgrid - Israel - Cloud Partners" all show **0 touches in DB**. The partner-ICP pitch is named in campaigns but not actually pushed at volume. | **A** | | 3 | "Marketplace-listing-attach offer is the new margin" (research-doc lines 38, 158, 184, 751-758) | **IGNORED** | Only campaign with "Marketplace" in name = `EC - Marketplace Credits Everything` (739 sends, 3.1% reply, 1 opp) — but actual targets are B2B service/dev agencies (Syno International, Weapp, WeDevelop, DX Heroes, Reach Digital) being pitched the standard *credits* hook, NOT marketplace-listing-attach. Avishay mentions "Marketplace integrations" exactly ONCE in 53 threads, as a passive service-list item (`avishay_threads.jsonl` thread 19d80e08d018083c). | **A** | | 4 | "MDF under-claim leaks ~40%" — pitch MDF-attach (research-doc lines 39, 270-274, 1392 "Angle 2") | **IGNORED** | Zero MDF-pitch campaigns. Only "funding" campaigns are `AccelX - AWS Funding Eligibility` (AWS Activate for startups, not partner-MDF) and `EC - Google Funding Credits Everything` (credits for startups). Avishay synthesis line 61: "MDF / AWS Activate / Azure Founders | 0 | Never mentioned." | **A** | | 5 | "Sovereignty/Nimbus is IL moat" (research-doc lines 36, 574-582, 918-927) | **IGNORED** | Zero campaigns with "sovereign", "Nimbus", "DORA", "BAA" in name or angle. Avishay synthesis line 60: "BAA / sovereignty | 0 | Never mentioned." Avishay-thread grep for sovereignty/Nimbus = 0 hits. Despite research-doc treating this as THE structural IL moat. | **A** | | 6 | "Azure Expert MSP has ZERO IL partners — gap opportunity" (research-doc lines 35, 114) | **IGNORED** | No campaigns targeting Azure-Expert-MSP-aspirant partners. The strategy doc explicitly flags "zero Israeli AEMSP" as a "gap opportunity"; ops shows zero engagement with that gap. | **A** | | 7 | "DoIT/AllCloud/CloudZone/Matrix/Terasky/2bcloud/Emind/Comm-IT/Bynet/Ness etc. are top targets" (research-doc lines 35, 100, 116-118, 130, 700-705) | **DISCONFIRMED** | Direct exact-domain check across 28 named IL/EU cloud-partner targets in `leads_snapshot.db`: **0 touches** at any of them. Only `rebura.com` exists in `companies` (1 person, 0 touches). DoIT, AllCloud, CloudZone, Matrix, Terasky, 2bcloud, Emind, Naya, Comm-IT, Bynet, Malam, One1, Automat-it, Caylent, Mission Cloud, SADA, Cloudreach, Cloudride — none touched. | **A** | | 8 | "Multi-cloud MSPs are highest-margin archetype" (research-doc lines 153-159, 600-601) | **IGNORED** | Only "multi-cloud" tagged campaign is `CloudCredits_05_MultiCloud` (202 sends, 1.0% reply, 0 opps) — and it's a credits campaign targeting startups, not multi-cloud-MSP partners. No persona segmentation in `campaigns_meta` flags multi-cloud-MSP as a target archetype. | **A** | **Composite score**: 1 VALIDATED, 5 IGNORED, 2 DISCONFIRMED — **research doc is ~12% executed**. --- ## Where The Gap Is Biggest (top 3) ### Gap 1 — Strategy doc says "sell TO partners"; ops sells "FOR partners (EC) to startups" Tier **A**. The research-doc ICP framing (lines 882-893) is: partners = GG's *customers*; GG sells outbound-as-a-service to 20-150 emp Premier-chasing partners. Reality: EC/WideOps is GG's only paying cloud-partner client (decay verdict shows 43 of 89 campaigns are EC/WideOps lifetime; `instantly_lifetime_stats.jsonl`); the rest of "cloud partner" sends are **outbound FOR EC to startups**, i.e. EC's end-customer prospects. The five named-but-empty partner-ICP campaigns ("MSPs Acquisiton Test", "Cloud Reseller - Growthgrid", "IL Cloud Partners HeyReach", "Growthgrid - US Cloud Partners - AWS", "Growthgrid - Israel - Cloud Partners" — all 0 touches in DB; `leads_snapshot.db`) suggest these were *planned*, named, then never sent at volume. This is the largest single execution gap. ### Gap 2 — Strategy doc names ~15 IL Premier targets; zero touches at any of them Tier **A**. Exact-domain audit of `leads_snapshot.db` against research-doc named targets shows 0 touches at DoIT, AllCloud, CloudZone, Matrix, Terasky, 2bcloud, Emind, Naya, Comm-IT, Bynet, Ness, Malam, Automat-it (research-doc lines 100, 116-118, 130, 700-705 — every Premier-named IL firm). The "IL wedge" the doc rates 8/10 in the scorecard has zero documented outreach to its named target accounts in the last 30 days of `touches` data. Caveat: `touches` covers Apr 2026+ only (decay verdict caveat line 3). Pre-April outreach to these accounts could exist in `replies_full` or instantly history, but the campaigns with names suggesting they targeted these firms (above) all show 0 touches even in the post-Apr window. ### Gap 3 — Sovereignty/Nimbus/Marketplace-attach/MDF/AEMSP — every "moat" angle is unexecuted Tier **A**. Research doc identifies five structural-moat angles unique to IL or unique to the niche: (a) Project Nimbus + Israeli sovereignty (lines 574-582); (b) MDF-claim-ready agency invoicing (lines 39, 270-274, 1392 Angle 2); (c) Marketplace-listing-attach for ISVs (lines 38, 184, 751-758); (d) Azure Expert MSP gap (lines 35, 114); (e) Tier-renewal-panic sprint pricing (lines 381-393). **Zero of five** have a corresponding campaign in `campaigns_meta` or any mention in `avishay_threads.jsonl`. Every "differentiation moat" is a paper moat. --- ## Research-Doc Blind Spots (what the doc DIDN'T anticipate) ### Blind spot 1 — The "credit-zombie" problem Tier **A**. Avishay's Feb 1 Fathom call (`118740894`, captured in `04_avishay_synthesis.md` line 48, line 90, line 190) flags that credits is the highest-traffic angle but it attracts "credit whores" — prospects who book the meeting, take the credits intro, then ghost. The research doc treats credits as a partner-economics topic (lines 86, 88, 154, 163-170) and never connects it to the *lead-quality death spiral*: credit-attraction lures wrong-fit prospects that drag the funnel. This shows up directly in `02_decay_verdict_by_campaign.md`: `Europe Cloud - Credits - Apollo List` peaks at 2.41% strong_prr in March, drops to 1.18% in April, verdict = fading. The doc had no theory for why credits-led ops would self-poison; Avishay diagnosed it in the room. ### Blind spot 2 — Copy drift inside the same "angle" Tier **A**. `06_credits_angle_autopsy.md` proves the credits angle's WIN-RATE dropped from 63% (Jan) → 17% (May) because the COPY drifted — not the audience. Specific phrase-level evidence (`06_credits_angle_autopsy.md` line 47 table): "do you use AWS or Google Cloud?" opener fell 48% → 0% Jan→May; "$50-100k" specific amount fell 27% → 0%; replaced by "premier-tier", "as a CEO", "30% off cloud bill", "$350k credits", P.S. footers. The research doc's `Section 11 — Messaging Angles That Land` (lines 895-904) names angles abstractly ("tier renewal panic", "MDF surplus", etc.) but assumes copy-execution is invariant. Reality: same-named angle can drop 4× in win rate within a quarter purely from copy entropy. This is the dominant fragility of the niche, missing from the doc. ### Blind spot 3 — Avishay-driven thrash overrides data Tier **A**. `04_avishay_synthesis.md` lines 156-211 documents 9 explicit contradictions where Avishay reversed direction within 30-90 days (Israel killed Feb → IL wedge to be added; quality preference Feb → 14 meetings/month demand April; cost-savings rejected Nov → credits demand April; geo data approved Apr 9 then immediately overridden with Cyprus/Baltics anecdotes). The research doc's `Section 17.5` calendars a re-decision at month 9 — but in actual ops Avishay forces a partial re-decision every ~30 days. The 90-day-monitor cadence the doc prescribes (lines 1637, 1741) is structurally incompatible with the anchor client's revision rate. Bonus: doc never anticipated that the anchor client's positioning ("save money" vs "make money" — wa_msg_index 26/56/161/179 in synthesis) would still be unresolved 6 months in. --- ## Three Strategic Moves That Would Close the Gap ### Move 1 — Actually send to the 15-firm IL Premier-named list (the IL wedge the doc rates 8/10) Tier **A** justification. The research doc's only 8/10+ scorecard dimension is "Israel wedge" (line 1725). Operationally it's untested: 0 touches at the 15 named IL Premier firms in `leads_snapshot.db`. The doc's *Condition 4* (lines 1659-1664) explicitly says push 1,500 Hebrew sends into IL cloud partners as the proof-of-wedge test — this has not been run. The IL Play 1 Hebrew campaigns DID run (`IL Play 1 - CEOs Hebrew 2026-04-23` 557 sends 0.9% reply 2 opps; `IL Play 1 - GTM Hebrew 2026-04-23` 306 sends 0% reply) but those targeted Israeli startup CEOs/GTM, NOT Israeli partner-firm Alliance VPs. Concrete: build a 60-person Apollo+manual list across DoIT, AllCloud, CloudZone, Matrix-Cloud, Terasky, 2bcloud, Emind, Naya, Comm-IT, Bynet, Automat-it, Ness, Cloudride, Tikal, Malam (`cloud-partner-niche-research.md` lines 700-705) targeting VP-Alliances / Head-Marketing personas with a "tier-renewal-sprint" pitch (doc lines 381-393). This is the cheapest A/B against the doc's Condition 4. ### Move 2 — Pivot the unspent "Premier Partner Credits" / "MSP Acquisition" campaign shells from credit-attraction to MDF-attach Tier **B** justification. Three campaigns are named to suggest partner-targeting but execute as credits-to-startup ("GG - Premier Partner Credits - EU DM - 2026-04" 998 sends 1.9% reply 0 opps; `MSPs Acquisiton Test` 0 touches; `Glassix - MSPs UK` 0 touches). The research doc's `Angle 2` (line 1392) recommends "MDF-eligible outbound for EU cloud partners" — invoice-structure that's MDF-claimable. Combined with the credit-zombie blind-spot diagnosis above, pivoting these shells from "credits to startups" → "MDF-claim-ready outbound for your EU partner firm" hits a non-zombie pool (partner VP-Marketing controlling MDF burn). Even one closed retainer at $6-12K/mo (doc lines 946-957) is meaningfully accretive to the current $2,500/mo EC retainer math. ### Move 3 — Stop drift: lock the Jan-Feb-Mar credits-copy template as the *only* approved variant; A/B from there Tier **A** justification. `06_credits_angle_autopsy.md` proves the OLD credits copy (Jan: "do you use AWS or Google Cloud?... we're a certified partner... $50-100k... worth a look?") still wins when it runs (`Europe Cloud - Credits - Apollo List` 39% W-rate in April even on a fatiguing list). The May variants ("premier-tier / 30% off cloud bill / as a CEO" — `EC_Cloud20K_2026-05`, `EC - 2026-05-03 - A1 LLM/GenAI`) hit 0%. The fix is a process gate not a strategy decision: any new campaign must reproduce the 5-element Jan template (concrete $ amount, AWS-or-GCP question opener, certified-partner frame, extend-runway value, short-CTA) before testing variants. This is the highest-leverage zero-cost move available; it requires no new lists, no new angles, no Avishay re-alignment — just refusing to ship copy that drifts. --- ## Inline 7-line summary 1. GG is hitting the research doc's 60/40 cloud-partner send-volume target (61.9% actual) — that headline-level alignment is real. 2. But every named structural-moat angle (Sovereignty/Nimbus, MDF-attach, Marketplace-listing-attach, AEMSP-gap, multi-cloud-MSP, named-IL-Premier-targets) is unexecuted; 12% of the doc is shipped. 3. The "sell TO partners" framing of the doc is structurally mismatched with the operational reality of "sell FOR EC to startups" — campaigns named after the doc's ICP are mostly empty shells in the DB. 4. The doc completely missed Avishay's credit-zombie problem (Feb 1 Fathom diagnosis) and the copy-drift collapse that turned the credits angle from 63% to 17% W-rate Jan→May. 5. The doc's 90-day re-decision cadence (lines 1637, 1741) is incompatible with Avishay's actual 30-day flip rate (9 documented contradictions in 6 months). 6. The IL wedge — the doc's only 8/10 dimension — has 0 touches at any of the 15 named Premier firms it identifies as targets. 7. Three highest-ROI moves: (a) actually send to the named IL Premier list, (b) repurpose the empty "Premier Partner / MSP Acquisition" campaign shells from credit-zombie attraction to MDF-attach pitches, (c) revert copy to the Jan template and gate all variants on phrase-level fidelity to it.