Cold Acquisition Strategy
August to November 2026. Three channels, one number worth learning, and a target that has to be deleted before anything else works.
realistic ceiling
instrumentation only
weeks 1 to 4
synthesized
The 50 to 100 calls per month target is wrong by 3 to 4x. Delete it.
Not defer it, not phase it in. It is a month 9 to 12 number at triple the budget, and even then Tom's calendar cannot hold it. The budget is correctly sized. Only the target is wrong.
Two ceilings, same answer
Media cost and Tom's calendar independently cap the engine at 25-40 booked calls per month, and they agree by coincidence rather than by design. ₪2,000 at Israeli Meta rates buys 15-35 booked calls, not 100. Reaching 100 costs roughly ₪7,500 per month, and it is 100+ hours of Tom system time against about 45 hours of real monthly slack.
"It pays for itself on consults alone" is false
That claim requires a cost per booked call at or under ₪37.50, which is unreachable in this market. Full funnel breakeven is about ₪112.50 per booked call, which is achievable, but only once product sales are counted in the numerator.
Builds are the business case
Builds are not upside sitting on top of a self funding consult engine. They are the entire economic argument. The metric that decides whether this worked is builds per quarter, not calls per month, and the ₪250 consult is a ₪231 per hour qualification filter rather than a profit centre.
Where the engine math breaks, and how it is fixable
The engine math does not hold, and it fails in a specific, fixable way. The budget is correctly sized. The target is wrong by 3 to 4x. Two independent ceilings, media cost and Tom's calendar, both land in the same place at 25-40 booked calls per month, and they agree by coincidence rather than by design. ₪2,000 at Israeli Meta rates (median CPM around ₪27, CTR 1.0-1.5%) buys 15-35 booked calls, not 100. Getting to 100 needs roughly ₪7,500 per month. Separately, 100 booked calls is more than 100 hours of Tom system time once scheduling, 40% no-shows, prep, notes and follow-up are counted, set against roughly 45 hours per month of real slack while GoMobile build #2, the website build and Monesys are all live. 50-100 free calls per month is not a target to defer. Delete it.
"It pays for itself on consults alone" is flatly false. That requires a cost per booked call at or under ₪37.50, unreachable in this market. Full funnel breakeven is about ₪112.50 per booked call, which is achievable, but only by counting product sales. So builds are not upside. Builds are the entire business case, and the metric that matters is builds per quarter, not calls per month.
Two more corrections to the stated model. The 25% free to ₪250 rate is a warm and referral number. Cold converts at 12-18%, and if Tom blends cold and warm in one bucket the referrals will mask a failing paid channel for months. And 7.5 minutes is talk time. Loaded cost per held call is about 20 minutes, and per booked call about 1 hour.
The binding constraint is not budget and not lead supply. It is Tom's non-delegable hours. Every channel that actually works at this size (phone calls to partners, answers posted in his own voice, the calls themselves, reviewing every artifact before it ships) requires Tom personally. Agents can do prep, drafting, list building and instrumentation. They cannot do the four things that produce the calls. Tom has about 8-10 hrs/week for acquisition while two client builds are live, and half of that goes to the calls themselves. Every pick below fits inside 4-5 hrs/week of Tom's channel work.
The calendar is the second constraint, and it is unusual. Israel is functionally impaired from 9 August to 4 October: school vacation, then Rosh Hashanah (11 Sept), Yom Kippur (20-21 Sept), Sukkot (25 Sept to 2 Oct), Simchat Torah (3 Oct), and נדבר אחרי החגים as a national deferral reflex. But this cuts asymmetrically. It kills operating business selling and leaves the idea stage audience, which is Tom's primary volume audience, more reachable rather than less. August and the chagim are for idea stage creative and machine building. October is for spend.
| Month 1 · Aug | Month 2 · Sept | Month 3 · Oct | Steady state · M6 | |
|---|---|---|---|---|
| Media spend | ₪600 | ₪900 | ₪2,000 | ₪2,000 |
| Free calls booked | 6-12 | 14-22 | 28-35 | 30-40 |
| Free calls held | 4-8 | 9-15 | 20-24 | 22-28 |
| ₪250 consults | 0-2 | 1-3 | 3-5 | 4-7 |
| Products (₪1,500-2,500) | 0-1 | 0-1 | 1-2 | 1.5-2.5 |
| Engine revenue | ₪0-2,500 | ₪250-2,750 | ₪2,750-5,250 | ₪3,500-7,000 |
Month 1 is instrumentation and will look like failure. That is the correct outcome.
Three channels for weeks 1 to 4
One paid channel that scales without Tom's hours, one free channel with the fastest time to first call on the list, and one partnership that arrives pre qualified at the paid tier. A fourth is built now and started on 5 October.
Meta Instant Form to WhatsApp thread to booked call
WhatONE campaign (Leads objective), ONE ad set (Israel, ages 25-58, all genders, Advantage+ audience, all placements including Reels and Stories), SIX ads inside it: three written to idea stage, three to operating businesses. Destination is a Meta Instant Form set to "Higher intent" with three fields only: name, a manually entered WhatsApp number, and one required open text field, במשפט אחד: מה אתה בונה או מנסה לתקן? The Meta lead webhook feeds the existing n8n flow, into the Notion Leads DB, then a WhatsApp ping to Tom's self chat. Tom replies from his own number within minutes and books the call inside the chat.
Why it beats the alternativesIt is the only channel that scales without Tom's hours. Israel is one of the cheapest quality Meta markets on earth (CPM around ₪27 against roughly ₪63 globally), and August to September is its seasonal trough. Tom's primary audience, someone with an idea in their head, types nothing into Google. That demand has to be created, and a free 7 minute call is the ideal offer for interruption media. Instant Forms report cost per lead natively with no pixel, no GA4 and no site instrumentation, which dissolves the tracking problem instead of waiting on it. It also removes the landing page, which is where this funnel loses 96-98% of clicks. The required open sentence halves volume and doubles quality, and that is the point.
CadenceLaunch 11 August. Agents check the account daily and report Sunday and Wednesday. Tom reviews creative once a week for 30 minutes. One change per 7 days, in this fix order: form question, then headline match, then creative angle, then audience.
WhoAgents build, monitor, qualify in WhatsApp, and report. Tom writes nothing and manages nothing except the reply-and-book and the weekly creative call.
Substantive answers in Israeli founder communities
WhatFive named Facebook groups: יזמים (~13.5K), יזמים ישראלים (~12.8K), הגיקיה (~15.7K), שאלות ותשובות – עזרה לסטארטאפים, and Innovation Israel (~11.3K), plus קהילת אוטומציה עסקית on prog.co.il, plus XPlace as passive intake. Agents run keyword monitoring (כמה עולה / לבנות אפליקציה / אפיון / אוטומציה / מערכת ניהול) and drop 5-8 live threads with drafted answers into Notion each morning. Tom edits and posts from his own account. Rules: answer completely and for free with a real number or a named mechanism; no link, no CTA and no "DM me" for the first three weeks in any group; after that, one soft line maximum. On freelance marketplaces, never bid to build. Bid the counter position: אני לא מציע לבנות עכשיו. יש לך 6 הצעות מחיר שאי אפשר להשוות ביניהן כי אין אפיון.
Why it beats the alternativesFastest time to first call on the entire list at 3-10 days, zero spend, zero legal exposure, and the intent is already declared. These people are literally typing the question the ₪250 consult answers. It is also the only channel where "no testimonials, no named clients" is irrelevant, because the answer itself is the proof. And it is the one that works during the chagim, when idea stage people are scrolling. Israeli groups are about to be flooded by graduates of the six or more Hebrew "build AI agents without code" courses, one cohort starting 26 August. The only way to be distinguishable from them is to be the person who gives away the answer they cannot give.
CadenceRead only in week 1. From week 2: 3-4 answers per week on Sunday, Tuesday and Thursday, about 30 minutes per answer including Tom's edit. The floor is 2 per week, not a target of 4. This is the item most likely to be dropped when delivery spikes, so the floor is set low enough to survive a bad week.
WhoAgents monitor and draft. Tom edits and posts. Non-delegable.
Tnufa grant consultant partnership
WhatAgents build a list of 25-30 Israeli firms and solo יועצי מענקים who shepherd Tnufa applications (search ליווי מענק תנופה and ליווי רשות החדשנות: AccountX, EKW, Cohen Mor, Outstand, Ofek Tzmiha, hacpa.net, plus the LinkedIn long tail). Tom phones or LinkedIn DMs 10 in week 1 and 10 more in week 2. The offer: he writes the technical section of their applicants' files, meaning the R&D work plan, milestones, technical feasibility and development budget breakdown, free for the first three files, at zero percent of their fee.
Why it beats the alternativesIt is the only channel where a third party is already assembling Tom's exact buyer, an Israeli individual with an idea, no company, no product and documented intent to build, and attaching ₪200k of funding to them. Tnufa pays up to 80% of a ₪250k budget to a pre company individual, and the eligible cost list explicitly names technical consultants. That single fact turns the ₪1,500-2,500 Viability Assessment from a discretionary purchase into a required artifact, and turns a ₪60k build into ₪12k out of pocket. The grant consultants are financial and administrative people who structurally cannot write a credible R&D milestone plan, so "free for three" reads as gap fill rather than desperation. These leads skip the free call entirely and enter at Assessment tier, which is exactly why the low volume does not matter.
Cadence20 approaches across weeks 1 and 2 (Tom, about 1.5 hrs/week). Then one maintenance touch per week. Ship the supporting page, יש לך רעיון? המדינה עשויה לממן 80% מהאב-טיפוס, by end of week 3 so partners have a URL to forward instead of a description.
WhoAgents build the list and the page. Tom makes every call. This is a peer offer and it cannot be delegated.
Job ad phone outbound
Agents build the list now at zero Tom time: a daily crawl of AllJobs, Drushim, JobMaster and company career pages for operational debt vocabulary (שליטה מלאה באקסל, VLOOKUP, הזנת נתונים, מעקב אחר הזמנות, הפקת דוחות, פריוריטי/חשבשבת/ריווחית) in roles like רכז/ת תפעול and בקר/ת מלאי. Discard every agency posted ad. Tom starts dialling 5 October, two 90 minute blocks per week (Tuesday and Thursday, 10:00 to 11:30), off the company's own published number, which sits outside the spam law and outside the consumer marketing registry.
Why deferred and not dropped: it is the strongest channel for the operating business audience and the only one that reaches the מוסך/יבואן owner operator that LinkedIn cannot touch. But that audience is genuinely absent from 9 August to 4 October, and Tom has no spare hours in August. Building the list costs him nothing. Dialling into a vacuum costs him the ramp.
| Rejected | Why, in one line |
|---|---|
| Cold email, any volume, any domain | Amendment 40 carries ₪1,000 per message in statutory damages, has no B2B carve out, and is a standing class action target. A quarantined domain is also 5-6 weeks of warm up for 2-5 calls per month. Skip entirely. |
| Bulk WhatsApp | Same law, plus it costs the number running Operator, client comms and Tom's personal life. Unrecoverable, for a ₪250 product. Write it as a hard AIOS rule on Monday. |
| Google Ads in month 1 | ₪400-700 buys 35-60 Hebrew clicks, which is statistically mute. Revisit on 5 October with the winning Meta hook as the free headline. |
| LinkedIn Ads | ₪1,500 buys roughly 85 EMEA clicks. Dead at this budget. |
| Performance Max | Eats ₪700 on garbage placements before a report is readable. |
| Optimizing Meta for the booked call event | The learning phase needs about ₪21,000 per month at a ₪100 CPA, so he never exits it. Optimize for the form lead and judge on raw cost per booked call. |
| Splitting the budget across two platforms | Both starve and both say nothing. |
| A LinkedIn posting and DM engine as a standalone channel | Competes for the same 2 hrs/week as community answering, with 2-3 weeks to first call against 3-10 days. LinkedIn is a delivery mechanism for partner recruitment here, not a channel. |
| Blog, newsletter, podcast, long form YouTube, Twitter/X, Medium, Instagram grid, Hebrew SEO agency | Zero calls inside 90 days. Hebrew commercial volume in this niche is thin, and English head terms are unwinnable for a young domain. |
| A daily or 3x weekly short form video schedule | Stacked on two live client builds, this is how the plan dies in week three. Clips only as a byproduct of ad creative. |
| Israeli Meetup groups | Dead. Founders Nation Israel: 1,141 members, last event May 2023, one attendee. |
| Accelerator portfolio companies | Funded and past the "should I build this" question. Their rejects are interesting, so queue that for month 3. |
| Referral fees to intermediaries | 10% of ₪250 is ₪25 and it insults the recipient. The free call is the gift. Money never leads. |
| Fiverr and Upwork as an acquisition strategy | 15-20% commission against a global supply pool, and it actively undercuts the ₪250 positioning. |
| English language paid acquisition | About $5.34 average CPC, so ₪1,500 buys roughly 75 clicks. International arrives free via the demos or not at all. |
| Pointing paid traffic at the email gated sample doc as the primary offer | Adds a 2-3 week nurture lag to a business that needs a signal in six weeks, and an email is a higher friction ask than a WhatsApp thread in Israel. The doc stays as the pre call asset and the retargeting offer. |
| Buying followers, fabricated proof, naming clients | Off the table by policy, and by law. |
Weeks 1 to 4, on the Israeli week (Sunday to Thursday)
Build the rails. Spend ₪0.
2-6 August
Monday morning, 3 Aug. Tom's own two hours, in this exact order.
- Meta Business portfolio, then Page, then Instagram, then ad account. Set currency ILS and timezone Asia/Jerusalem at creation. Both are permanent, and getting either wrong means abandoning the account and rebuilding. Add Tax ID 302148911 to billing so VAT is charged correctly and stays reclaimable. This is first because new accounts get reviewed and throttled for 24 to 72 hours.
- GA4 property, linked to Firebase
octomonic-e5f37, measurement ID into the scaffolded slot insrc/lib/pixels.js. - Search Console. Verify octomonic.com, submit the sitemap, check Hebrew and English properties separately. The clock has to start now because the data accumulates slowly.
- Create the Dataset (pixel) and paste the ID into the scaffold.
Tom, rest of week (about 2 hrs)
- Kill the manual approval step. Booking becomes instant confirmation into a real slot. Set the hard cap at 2 slots/day, Sunday to Thursday, one contiguous block, which is a ceiling of 44 booked per month and about 26 held. Enforce it in the tool, not by willpower.
- Dial the first 10 Tnufa grant consultants on Tuesday and Wednesday.
- Join the 5 Facebook groups and prog.co.il. Read only. Post nothing this week.
Agents, all week
- Ship OG share cards on all 10 demos. Verified broken:
grep -l 'og:image' dist/demos/*/index.htmlreturns zero of ten, so every demo link Tom pastes anywhere currently renders as a dead grey card. Screenshot each at 1200x630, composite the branded strip plus name plus a one line descriptor, write og:image, og:title, og:description, twitter:card and og:locale into each index.html, and fix the two demos whose<title>reads as an internal name. Half a day, and the highest ROI item in the plan, because every other channel ends with someone clicking a demo link. - 4 or 5 real Page posts plus profile and cover from
assets/branding, so the Page is not empty when someone clicks the advertiser name. - n8n: instant booking confirm, WhatsApp reminder at T minus 24h and T minus 2h with one tap reschedule, sample viability spec auto sent 24h before every call, a
sourcefield plus free text איך הגעת אליי? on the booking form, and UTM capture into Notion Leads with a fixed vocabulary (meta,community,partner,jobad,referral). - Meta Conversions API node on the existing
api-brain.octomonic.comflow.
Copy changes, shipped this week (free, and every lens converged on them)
- Name the free call once and use the identical string everywhere: שיחת 7 דקות and "the 7 minute call." A named thing reads smaller than "a quick call."
- Rebuild the ₪250 as a named product with a written output: a 30 minute session that ends with a one page written verdict within 24 hours covering what you are actually building, whether it is viable as described, the three things that will break it, real cost and timeline, and what to do next including "do not build this."
- Publish the credit forward rule: the ₪250 is credited in full against the Viability Assessment or Workspace Setup if taken within 30 days. It costs nothing and removes the "paying twice" objection.
- Change the free call's stated job. It is a fit check to decide whether the paid session is worth either party's time. Not a mini consult, and not a sales call.
First spend, first answers.
9-13 August
- Agents: write 6 ads, Hebrew primary with English duplicates. Three idea stage: the price anchor; יש לך רעיון למערכת? עדיין לא צריך מתכנת. צריך מישהו שיגיד לך אם שווה לבנות את זה; and לפני שמשלמים 60,000 ₪ על פיתוח: לדעת בדיוק מה מזמינים. Three operating business: אם העסק רץ על אקסל וזה כבר לא מחזיק; the demo and proof angle; and the one question talking head. Camera optional path: the identical script as ElevenLabs VO over screen recorded demos, so the on camera question never has to be answered.
- Agents: Playwright click paths through all 10 demos, producing 8 to 12 vertical clips at 1080x1920, 6 to 12 seconds, silent first, with burned in Hebrew captions and a branded end card. This is the entire creative library for Reels and Stories, at zero cost.
- Launch Tuesday 11 August: ₪30/day, one campaign, one ad set, six ads, Instant Form destination. Idea stage creative weighted 60/40, because operating businesses are on vacation.
- Tom: first 3 community answers. No link, no CTA.
- Tom: 10 more partner approaches. Pick one second track, either bookkeepers or web studios, not both.
- Agents: build
/he/startand/en/start. No nav, no footer, a headline repeating the ad hook verbatim, a muted demo loop above the fold, the slot picker embedded inline rather than linked, and a floating WhatsApp button. Every ad URL uses the explicit language path with UTMs and never the bare root, because the SPA drops UTMs on client side nav (already documented inprojects/octomonic-website/STATUS.md).
First calls held.
16-20 August
- The pre call mock goes live. Booking collects a two sentence idea description, agents spin up a rough clickable mock of that specific idea, and Tom opens the call sharing their idea on screen. This is the single strongest differentiator available. It converts the agent team into the offer rather than into more content about the offer, it is the strongest possible answer to having no testimonials, and it makes the call very hard to no-show because something was made for them. Every mock is also a candidate demo and a candidate post, so the content pipeline's marginal cost goes to near zero.
- Kill any ad at ₪150 spent with zero leads. No exceptions, no sentiment.
- Tom: 3 to 4 community answers. First DMs should land this week or next.
- Agents: start the job ad crawler (zero Tom time, ready for 5 October).
- Agents: ship the Tnufa eligibility page with the 6 question self check and the free call CTA framed as עשר דקות כדי להגיד לך בכנות אם בכלל שווה להגיש. Say plainly on the page that Tom does not file the application. He writes the technical half and builds the thing.
First read, first decision.
23-27 August
- Read the four numbers. About ₪600 spent should have produced 8 to 13 form leads, then 5 to 8 booked calls, at a cost per booked call of ₪60 to ₪130.
- Add
/he/startas a second destination inside the same ad set, head to head against the Instant Form. That is the one A/B worth paying for. - Tom, one focused day: the מאגר היועצים והמרצים application at sba.org.il (certificates, CV, Osek Murshe docs, the specialisation form under digitisation, technology and operations, plus the lecturer flag). Then email BDO and Aviv Business Consulting to say he is registered. Fire and forget. It answers in 2 to 4 months or it does not, and it is the only route to institutional third party credibility under a no testimonials policy.
- Tom decides the anchor question (see below).
- Week 4 gate: cost per booked call under ₪150 and show rate above 60% means September proceeds at ₪900. Over ₪200 means do not increase spend. Change the form question and then the landing headline first, in that order.
In Israel ₪5,000 is a small project, not an hour of judgement. The honest comparable for senior consulting is ₪400 to ₪800 for a first meeting. A claimed 20x discount triggers "what's the catch," and worse, the anchor sits above Tom's own flagship ₪1,500-2,500 deliverable, which quietly tells the buyer the spec cannot be the real thing.
The fix keeps the positioning intact and changes only what the number points at: anchor against the cost of building the wrong thing. A full specification from an Israeli product shop is ₪10,000-16,000. A medium app build is ₪30,000-150,000. The ₪250 session exists so you find out for ₪250 whether to spend the ₪30,000. Every number is publicly checkable, it makes ₪1,500-2,500 look like a 6x saving rather than a thin document, and it reframes the purchase from cheap advice to cheap insurance.
Run both versions as ad copy from week 2. It costs nothing to test and the answer arrives in three weeks. Tom's call, but it is the highest value copy decision in the plan.
₪3,500 over 90 days, deliberately back loaded
Sequenced against the Israeli calendar and held under the ₪1,000-2,000 per month envelope. Roughly 60% of the total is withheld until 5 October.
| Window | Spend | Daily | What runs |
|---|---|---|---|
| 11-31 Aug | ₪630 | ₪30/day | Meta Instant Form, 6 ads, idea stage weighted 60/40. Labelled instrumentation, not lead gen. |
| 1-10 Sept | ₪500 | ₪50/day | The clean window. Full 6 ad rotation, both audiences, both destinations. |
| 11 Sept to 4 Oct | ₪400 | ₪17/day | Chagim. Idea stage creative only, zero operating business. Low and steady, because a paused ad set restarts learning while a reduced one keeps its history. |
| 5-31 Oct | ₪2,000 | ₪60/day + ₪250 retargeting | Full push. The retargeting ad set (demo viewers and spec downloaders, offering the sample document) starts once the audience crosses about 500 people. The job ad phone track starts 5 Oct at ₪0. |
The market lens said zero paid spend in August. The paid lens said launch now. Both are half right. Waiting until 1 September means 7 clean days of data, then three weeks of chagim noise, then a restart on 5 October, which puts the first trustworthy cost per booked call in mid November. Too slow. Spending ₪630 in August on idea stage creative only buys the calibration number early, exploits the Israeli CPM trough, and targets the one audience that is more available during school vacation rather than less. The operating business creative stays throttled until 5 October.
A quarantined cold email domain (₪150-250/month): ₪0.
Sales Navigator: ₪0 until DM to call clears 4%, which is a month 3 question at the earliest.
ZenRows for list building: that account is registered to gpi@octomonic.com and holds the client's vendor receipts, so using it for OctoMonic prospecting is cross billing a client's tool. Use the Apify free tier, Firecrawl, or plain requests.
- 01
Kill the losing audience's creative first
If idea stage carries a materially lower CPL, kill the three operating business ads and put 100% behind the winner. Costs nothing, and it is typically worth 30-40%.
- 02
Kill the losing destination
If
/he/starthas not beaten the Instant Form on cost per booked call by week 6, stop paying for it and run 100% Instant Form. - 03
Reduce, do not pause
Drop from ₪60/day to ₪35/day rather than pausing, because pausing destroys the signal pool.
- 04
Hard stop
60 live days, ₪3,000 cumulative, and zero paid ₪250 consults means kill paid entirely. There is no second paid channel worth ₪2,000 at this size. The money goes back in the pocket and the hours go to the job ad phone track, which costs ₪0.
Six metrics, weekly, one automated line
Reported Sunday morning as a single WhatsApp line. Do not report or decide on monthly revenue until month 4. At roughly one product per month the revenue signal is pure noise, and any scale-or-kill decision in months 1 to 3 based on it is a coin flip dressed as analysis.
| # | Metric | Target | Amber | Kill or reallocate |
|---|---|---|---|---|
| 1 | Cost per booked call (cold only) | ≤₪120 green ≤₪80 | ₪120-200. Fix in order: form question, headline match, creative, audience. One change per 7 days. | Above ₪200 after ₪2,000 cumulative and 45 live days, kill Meta. |
| 2 | Show rate | ≥70% | 50-70%. Tighten the reminder ladder and ship the pre call mock. | Under 50% sustained over 20+ bookings means the offer is attracting tourists. Add friction to the form. |
| 3 | Free call to ₪250, cold and warm tracked separately | 12-18% cold | 8-12%. The ₪250 is not reading as a product, so rework the named deliverable. | Under 8% across 20 or more cold calls is an audience quality failure. Cheaper clicks will not save it. |
| 4 | ₪250 to product | ≥25% | 15-25% | Under 15% across 8 or more consults means the Assessment scope or price is wrong, not the channel. |
| 5 | Tom hours per booked call | ≤1.0 | 1.0-1.5. Move more prep to agents. | Above 1.5 means cut the calendar cap before cutting spend. |
| 6 | Builds per quarter (the real one, lagging) | ≥1 by 31 Dec | None | Zero by 31 Dec with 8 or more products sold means the ladder is broken above ₪2,500, not below it. |
Six failure modes, in the order they are likely to arrive
- 01
Tom's own hours. Most likely, weeks 3 to 5.
Community answering is the first thing dropped when GoMobile #2 or the website build spikes, and once dropped it does not restart.
FixThe floor is 2 answers per week, not a target of 4. Agents pre draft into a Notion queue so posting is a 10 minute edit rather than a 40 minute research task. If the floor breaks twice in a row, cut the calendar cap to 1 slot per day rather than letting call load and delivery collide. - 02
No-shows. Weeks 2 to 4.
Free plus cold plus stranger is a 40-50% no-show environment. There is no published benchmark for this exact case, and the honest planning number is 40%.
Fix, already in week 1Instant confirmation, T minus 24h and T minus 2h WhatsApp with one tap reschedule, the sample spec 24h before, and from week 3 the pre call clickable mock. Cutting no-show from 50% to 25% is worth the same as a 33% budget increase, at zero marginal media cost. - 03
Message-market fit. Weeks 2 to 3.
Expect 4 of the 6 hooks to be dead.
FixKill at ₪150 with zero leads, no sentiment. If all six produce a CPL above ₪80, the problem is the offer and not the creative. That is the moment the build cost anchor replaces the ₪5,000 anchor and the free call gets renamed around the artifact rather than the conversation. - 04
Spend efficiency. Weeks 5 to 6.
Cost per booked call lands at ₪150-250 and stays there.
Fix, in order100% Instant Form, idea stage creative only, then drop to ₪35/day rather than pausing. Do not increase spend to "get out of learning." At a ₪100 CPA that needs about ₪21,000 per month and it is not available. - 05
Delivery collision. October, and by far the most expensive.
Tom books 35 calls in October and ships GoMobile build #2 late. This costs a real paying client and the referral flywheel that lives inside good delivery, in order to chase ₪250 sessions.
FixThe hard calendar cap, enforced in the booking tool from day one. When the block is full the site offers the next available day. If demand exceeds the cap, do not add hours. Raise the ₪250 or tighten the qualifying question. Overflow is a pricing signal, not a scheduling problem. - 06
The quiet one: measuring the wrong rung.
In month 2 the engine will look like it is losing money, because it is. ₪900 spent, maybe ₪500 in consults. The failure mode is killing it there.
FixThe pre committed thresholds in section 5, plus the standing reminder that the ₪250 consult is a ₪231/hr qualification filter rather than a profit centre, and the Assessment runs ₪100-208/hr against 12 to 15 hours of senior work. The money is one ₪25k+ build every 3 to 4 months. Judge on builds per quarter.
What each specialist concluded
Six independent lenses ran against the same brief. They disagreed on timing and agreed, without coordination, on the one thing that mattered.
Lens 01
Economics
The engine breaks in two independent places that happen to cancel out. Volume: Israeli benchmarks (Google Ads CPL ₪100-400 in SaaS and tech, ₪80-200 in real estate, ₪30-150 for courses) and bottom up Meta math (Israel median CPM $9.15, which is ₪27.45 at 3.00 ILS/USD, CTR 1.0-1.5%, LP conversion 2-4%) both land at ₪46-137 per booked call. So ₪2,000 buys 15-44 booked calls, not 100. Hitting 100 would need a 9-14% landing page conversion rate on cold social traffic, beyond any published benchmark.
Time: each booked call costs about 1.04 hours end to end, so 100 booked calls is 104 hrs/month on top of the roughly 153 hrs already owed to GoMobile, the website build and Monesys.
The lucky part: ₪2,000 buys roughly what he can actually absorb, about 25-40 booked calls, so the budget is correctly sized and only the target is wrong.
Lens 02
Paid
Meta first, alone, for month one, and not because Meta is better. Israel's Meta inventory is one of the cheapest quality markets on earth (12 month median CPM about $9.15 against about $20.93 globally), and Tom's primary audience types nothing into Google. There is no search query for "I have a vague idea and don't know if it's worth building." That demand has to be created, and a free short call is the single best suited offer type for interruption media.
Correct the target engine now: ₪1,500-2,000 per month buys 15-30 cold booked calls, not 50-100. Getting to 75 per month at a realistic ₪100 per call costs about ₪7,500 per month. The deliverable of the next 60 days is not volume. It is one number, accurate enough to bet a bigger budget on. Also correct the 25% free to ₪250 assumption: that is a warm and referral figure. Model cold at 12-15% and track it separately, or the blended number will lie for months.
Lens 03
Organic
Classic SEO is not the acquisition channel at this stage and Tom should stop thinking of it as one. Hebrew commercial volume in this niche is thin, English head terms like "automation consultant" are unwinnable for a young domain inside 12 months, and the traffic would not be Israel anchored anyway. A general blog, a newsletter, long form YouTube, a podcast, Twitter/X, an Instagram grid, Medium cross posting and a Hebrew SEO agency are all straight time sinks right now.
What is worth building is the narrow slice where Tom already owns an unbeatable asset, ten live clickable products, plus real numbers pricing content aimed as much at ChatGPT and AI Overviews as at Google, where competition is years behind. The channels that actually produce calls in weeks are unglamorous: substantive answers in Israeli founder groups, and LinkedIn run as a targeted DM engine with posting as the credibility backdrop.
Lens 04
Outbound
Classic cold outbound, meaning a sequencer, a bought list and 3,000 emails a month, is the wrong engine for this business in Israel, and not mainly for deliverability reasons. Amendment 40 to the Communications Law (חוק הספאם) makes any unsolicited commercial message worth up to ₪1,000 per message in statutory damages, with no proof of harm required and no B2B carve out, and it is a standing class action target. That kills mass cold email to Israeli recipients and kills bulk WhatsApp outright, which would additionally cost Tom the number running Operator, client comms and his personal life.
What works at one person plus agents scale: phoning owners off public hiring signals (a job ad demanding שליטה מלאה באקסל is a company publicly announcing it runs on spreadsheets), answering publicly declared asks in groups and marketplaces, and sending a company specific one page spec at a private URL instead of a meeting request, because the spec is the ₪1,500-2,500 product.
Lens 05
Partnerships
The best channel for this business is not a channel. It is a government grant. Israel's Tnufa track at the Innovation Authority gives individual entrepreneurs, pre company, up to ₪200k at 80% of a ₪250k budget, and the eligible cost list explicitly names subcontractors and technical consultants. The application requires a technical work plan that the grant writing consultants filing these are structurally weak at. That single fact turns Tom's ₪1,500-2,500 Viability Assessment from a purchase into a step toward ₪200k, makes a ₪60k build cost the founder ₪12k out of pocket, and hands him a partner class with a real problem he can solve for free.
On professional intermediaries, the honest answer is that the referrer is the bookkeeper, not the accountant. Bookkeepers sit inside the client's manual mess every month and benefit directly when it gets cleaner, while רו"ח are risk averse, refer almost nothing, and take a year to warm. Lawyers are worse.
Lens 06
Market
Free consultation is table stakes in this market, not a differentiator. A direct Israeli competitor (automaziot.ai, page updated 2026-08-02) offers שיחת אפיון חינם scheduled within 48 hours, no commitment, plus a written action document at the end, and explicitly claims אנחנו לא רק מייעצים, גם מיישמים. That is Tom's exact "senior judgement, and I actually build it" wedge, already taken verbatim by a shop with a Meta pixel, FAQ schema, LP funnels and a phone number.
The ₪250 price itself is genuinely well placed, sitting under a lawyer's ₪400-800 first consult and mid range for a ₪300-800/hr business consultant, and paid discovery benchmarks support it: a $50-150 fee is the documented sweet spot, cutting no-shows from 40% to under 10%. What will not survive contact is the ₪5,000+ anchor.
What only Tom can do, in order
Agents own everything else. These eight items are the plan. If none of them happen, nothing else on this page matters.
- 1
Monday 3 Aug, 08:00. Create the Meta Business portfolio, Page, Instagram and ad account.
ILS, Asia/Jerusalem, Tax ID 302148911. Identity bound, irreversible, and it gates everything else with a 24 to 72 hour review lag.
- 2
Monday 3 Aug. Set the calendar cap and kill the approval step.
2 slots per day, Sunday to Thursday, one contiguous block, instant confirmation. This is the single decision that protects the two live client builds.
- 3
Tuesday and Wednesday. Phone the first 10 Tnufa grant consultants.
Peer to peer, and it cannot be delegated to an agent or to an email. Ten calls, about 90 minutes total, and it is the highest value per hour in the plan.
- 4
From week 2. Post every community answer from his own account, in his own voice.
Agents draft. The credibility is non-transferable.
- 5
Take every free call and every ₪250 consult.
This is the product.
- 6
Approve every ad hook and every outbound artifact before it ships.
One wrong fact destroys the artifact channel permanently, and agents cannot verify their own claims.
- 7
Decide the anchor.
Keep ₪5,000+, or re-anchor on the cost of building the wrong thing. Test both in ads from week 2 and decide by week 4.
- 8
Say no to the 45th booking.
Every other item on this list is reversible. That one is not.
Delete the 50 to 100 target, build to 30-40 booked calls per month capped by the calendar, run exactly three channels (Meta Instant Form to WhatsApp, community answering, Tnufa grant consultants), spend ₪630 in August purely to learn what a cold booked call costs, hold 60% of the 90 day budget until 5 October, and judge the whole thing on builds per quarter, because the consults will never pay for it and were never going to.