Outbound Playbook v4 — Concrete Leveling Partner Offer
The week-of-Jul-29 outbound plan — what changed, the one provable number, the scripts, the list, and the daily block.
1 Start here — the decision
You asked for a playbook to run this week. Running it un-parks cold outbound, which has been formally parked since the Jul-18 focus directive. Treat this doc as that un-park, scoped narrowly.
What is un-parked: you calling contractors, leaving voicemails, and texting people who gave you permission on the call. What stays off: every automated send. auto_send stays FALSE, Jean stays in suggestive mode, the SMS ladder stays off. Nothing leaves the building unless you personally pressed send.
That split resolves the Jul-22 manual-mode conflict without needing a new ruling — your own hand-dialled calls and hand-sent texts were never automation.
The three things that make this week different from every prior outbound attempt:
- 1The offer changed. It is no longer "$0 down, 10+ appointments in 7 days or you don't pay." It is the partner model: you front everything, plug into their CRM, take 10% of attributed revenue.
- 2You finally have a real number. Eight days of live Dallas data. See section 2 — and note that it kills the number you were about to put in the copy.
- 3You have a gift that costs nothing to give. The demo bot builds a working version of their website with the bot installed, in about 20 seconds, from just their URL.
The honest constraint: cold email cannot go live this week. DKIM and DMARC are still missing on stpierre.ai (verified today), and warmup takes 2–4 weeks. This week is phone-first, text-second, with the email foundation started in parallel so it is usable around Aug-19.
2 The one provable number
Action #1 from the Jul-28 CL market map was "reconcile the $15/booked-lead claim." Done. Here is the live data.
Source: Meta ad account act_3289323034532141 (*Concrete Leveling), Jul-21 → Jul-29, cross-checked against sp_client_leads in the brain with test rows excluded.
| Measure | Value | ||
|---|---|---|---|
| Total spend, all 3 CL campaigns | $406.62 | ||
| Spend, `CL \ | Dallas \ | $100pd` alone | $328.30 |
| Messaging conversations started (Meta, 7d) | 24 | ||
| Real leads landed in the brain (Jul-22 → Jul-29) | 33 | ||
| Real qualified leads | 16 | ||
| Meetings actually booked with Dan so far | 1 |
What that makes provable:
- ~$25 per qualified concrete-leveling lead — $406.62 ÷ 16 qualified. This is the number to use.
- ~$17 per homeowner conversation started — $406.62 ÷ 24. Safe, and easy to say on a call.
- 33 real homeowner conversations in 8 days on about $50/day — a volume claim, no cost math needed.
The $15/booked-lead claim is dead. Do not use it. Cost per booked job today is $406.62 ÷ 1 = $407. The $15 figure has no support anywhere in the data. Per the per-niche ticket standard, an unknown stays "unknown" — never a default.
Caveats you should be able to say out loud if pushed:
- The 33 brain rows and Meta's 24 conversations are not a 1:1 match; different attribution windows and some rows arrive outside the ad path. That is why the conservative claim is the qualified-lead number, not the raw-lead number.
- This is 8 days in one metro for one client. Say "in Dallas, over the last week" — do not imply a long track record.
- Dan's outcomes so far: 1 meeting booked, 1 proposal sent, 2 engaged, 1 in his CRM, 2 declined, 1 out of area. That is an honest picture of a system that is working and young.
What you are actually selling with this: not the cost, the fact that you have a live dashboard with real numbers in a niche where not one competitor has published a single CL client result. First to publish real numbers owns the credibility in this market.
3 The offer, in your words
The old cold-call offer is retired. So is $1-start, the deposit tiers, the weekly tiers, and "10+ appointments in 7 days or you don't pay."
The partner frame:
"I'm not a lead vendor and I'm not a retainer agency. I'm a revenue partner. I front the ad spend, I build the system, I plug it into whatever CRM you already use — and I get paid 10% of what actually closes from it. If nothing closes, I don't get paid."
The common enemy is always the model, never a company. Shared-lead marketplaces that resell the same homeowner to four contractors. Retainers that bill whether or not the phone rings. Never name Angi, HomeAdvisor, LevelRight, or any person as the villain — attack the model.
The free thing you lead with: the chatbot. It lives on their website, answers homeowner questions, collects the intake, and hands the lead to their pipeline. You give it away. It is how every lane — direct offer, cold traffic, cold outbound, JV — opens.
Why this lands in this niche specifically: zero of roughly ten marketing players in concrete leveling sell leads or performance deals. Everything out there is a $7K website, $995/mo SEO, or a $17 PDF. The 10%-of-attributed-revenue model has no analogue anywhere in the set. You are a category of one, and you can say that without exaggerating.
Language worth stealing, verbatim from their world:
- "You still need to keep that unit busy every day, and that's going to take a lot of marketing." (What contractors say to each other on Reddit about the $355K-per-van pitch.)
- "50–70% less than replacement" — the price anchor every homeowner in this niche already knows.
- HMI tells every rig buyer to budget about 5% of revenue, roughly $48K/yr, for marketing — then hands them door hangers.
4 The call script
The full craft reference is the Cold-Calling Field Guide at 01-Projects/StPierre-GTM/outbound/st-pierre-cold-calling-field-guide.html — screenplay, objection turns, dispositions, delivery coaching, the call-review scorecard. What is below is only what changed for the partner offer. Everything else in that guide still stands.
The sequence: Permission → Problem → Offer → Observation → Low-stakes ask → Book.
Opener
"Hey [First name] — [First name], it's Akash St. Pierre. I know I'm catching you out of the blue. Can I take 30 seconds, tell you exactly why I called, and you tell me if it's relevant?"
Then pause. Do not fill the silence. The first 7–10 seconds only has to earn the next 30.
The 30 seconds — rewritten for the partner offer
"I work with concrete leveling contractors. Most of them don't have a lead problem — they've got a calendar that swings between slammed and empty, and they're still renting demand from shared leads or Google.
I do something different: I front the ad spend, I build the whole thing, and I only get paid a percentage of what actually closes. Right now in Dallas I'm running that for a leveling contractor — 33 homeowner conversations in the last 8 days, about $25 a qualified lead.
I noticed [specific fact about their business]. That usually means [business consequence]."
The ask — this is where the free bot does the work
"Before we even talk about any of that — can I just send you the free piece? It's a chatbot that sits on your website and books the homeowner while you're on a jobsite. Takes me about a minute to build you one on your actual site. No charge, no strings, yours whether or not we ever talk again. Can I text it to you?"
A yes here is the whole call. You are not asking for a meeting. You are asking permission to text. That is a much smaller yes, and it is what makes touch 2 legal and welcome.
If they want the meeting anyway
"Would you be completely opposed to 30 minutes Tuesday or Wednesday? I'll have your version built before we talk, so you can play with it first."
Then: mornings or afternoons → two specific times → send the invite while still connected → confirm it arrived.
Voicemail
"[Name], Akash St. Pierre. I looked at [Company]'s site and built you a free chatbot for it — books homeowners when you're on a job. It's yours either way, no charge. My number is [number]. Again, Akash St. Pierre."
Objections that changed under the new offer
| They say | You say |
|---|---|
| "What's the catch?" | "There's no upfront. I make money when you make money — a percentage of what closes from what I build. The exact terms are in the agreement; I'm not going to invent numbers on a cold call." |
| "So you're a lead vendor." | "Opposite. A lead vendor sells the same homeowner to you and three competitors. I build you your own thing, in your own CRM, and I don't get paid unless it turns into work." |
| "We already have someone." | "Good — I'm not asking you to fire anybody. Do you own the data, and do you know what a booked job actually costs you right now? If yes, this is just a comparison for the back pocket." |
| "Prove it." | "Fair. My deepest history is foundation and waterproofing. On leveling specifically, what I've got is 8 days of live Dallas data — 33 homeowner conversations, about $25 a qualified lead. I'll show you the dashboard. And the demo I'm sending you is your own site, so that part you don't have to take my word for." |
| "We tried Facebook." | "I wouldn't ask you to run the same campaign again. Was the breakdown lead quality, follow-up speed, or close rate?" |
The rule of two. Two real attempts after an objection, then you book, set a specific callback date, or exit cleanly. Do not argue with a no. DND means DND, forever, on every channel.
Permanently banned, no exceptions: fake-customer openers, "we were supposed to talk," competitor-callback voicemails, implied referrals, borrowed logos, invented scarcity, any claim of equal proof across niches. Territory exclusivity is real and checkable — that is the only scarcity you get to use.
5 The demo gift — how to actually send it
This is the piece that makes this week different from every prior outbound push, and it is already live.
Two ways to deliver it:
- 1Self-serve (fastest). Send them
stpierre-demo-bot.pages.dev. They type their own website URL, and about 20 seconds later they are looking at their own site with the bot installed and talking. No work from you. - 2Pre-built (for the accounts you care most about). Drop a config JSON in
01-Projects/SP-Demo-Bot/configs/, build, and they get a cleanstpierre-demo-bot.pages.dev/{slug}link. Use the existingsample-concrete-leveling.jsonas the template.
The text, after they said yes on the phone:
"Here's the thing I built for [Company] — it's your actual site with the bot on it, not a slide deck: [link]. Made it before you paid me anything so you can see it's real. Have a play with it. Reply STOP to opt out."
Rules that are not negotiable:
- No cold first-touch text. Touch 1 is always a call or a voicemail. The text is only ever the thing you got permission for.
- The SMS mirror — where the demo texts them a sample transferred lead — only fires if they type their own number inside the demo. Never fire it at a number you already have on file.
- Any reply or link click gets an answer within 5 minutes.
- Every text carries STOP. STOP suppresses instantly and forever, across every channel.
- Text only 9 AM – 7 PM their local time.
6 This week, day by day
Your maker block (5–8 AM) stays content-only. Outbound starts after it.
The daily shape:
- 18:50–9:00 — say the opener, the 30 seconds, and your three worst objections out loud. Every day. This is the highest-return ten minutes in the whole block.
- 29:00–9:20 — warmest first: callbacks, replies, anyone who touched a demo link.
- 39:20–10:40 — protected cold block. GHL on one screen, calendar on the other. No Slack, no inbox.
- 410:40–11:00 — dispositions, send the demos you promised, confirm bookings, pick one call to listen back to.
The week's targets — 15 verified attempts a day, five days:
| Target | Why this number | |
|---|---|---|
| Dials | 75 | 15/day × 5, across ~25 accounts at 3 attempts each |
| Owner conversations | 18–25 | Direct lines to small contractors connect far better than corporate lists |
| Demo permissions ("yes, text it") | 6–10 | The ask is tiny — this should be the easiest yes on the call |
| Booked calls | 2–3 | The actual scoreboard number |
These are hypotheses, not forecasts. After the first 100 attempts, throw them out and use your real rates. The field guide's diagnostic bands: below ~5% connect points at list or number health before it points at your script; roughly 8–10% conversation-to-book is a normal early baseline.
Wednesday checkpoint. Read the funnel per stage, not blended. If connects are weak, the problem is the list or the numbers. If connects are fine but nobody takes the demo, the problem is the 30 seconds. Change delivery first, then message, then list, then target — in that order, one at a time.
Two zero-cost moves to do in the gaps this week:
- Join the "Concrete Lifting World Wide" Facebook group. This is where CL contractors literally ask each other whether they should target homeowners directly. Lurk and learn the language.
- Follow Tyler Link (Concrete CEO), Chris Koerner, and LevelRight. Recon, not outreach.
7 Who to call
The Week-1 call sheet is built and sitting at 01-Projects/StPierre-GTM/dream900-activation-2026-07-23/week1-call-sheet-CL-CS-2026-07-23.csv. Every row carries a verified decision-maker, a direct phone, a written hook, a 30-second explanation, and a do-not-claim list.
24 rows. Use 17 of them.
Pull two rows before you dial. Foundation Strategies and Construction (San Jose) and DESKA Foundation Repair (Alexandria VA) both have hooks marked refuted by QA — the claimed gap isn't there. Calling with a refuted hook is worse than calling with no hook. Also treat the four weakened hooks (Pinnacle, Quality Foundation Austin, Master Foundation Knoxville, Grand Rapids) as "no hook" — call them, but lead with the offer, not an observation.
Texas first, because Dallas is where your proof is and you can say "I'm running this down the road from you right now":
| # | Company | City | Owner | Phone |
|---|---|---|---|---|
| 2 | Crown House Leveling & Foundation Repair | San Marcos, TX | Robert L Moss | +1 512-691-9009 |
| 4 | Above All Foundation Repair | Converse, TX | Clay Crain | +1 512-631-1300 |
| 5 | El Paso Concrete Leveling | El Paso, TX | — | +1 915-331-4225 |
| 6 | G.L. Hunt Foundation Repair | Fort Worth, TX | Hayden Slack | +1 512-234-3470 |
| 1 | Pinnacle Foundation Repair | Grand Prairie, TX | Robby Brown | +1 972-251-0018 |
| 3 | Quality Foundation Repair Austin | Austin, TX | Simon A. Wallace | +1 737-260-7340 |
Then the Oklahoma cluster (Oklahoma Foundation Solutions, Edens Structural, Level Home Foundation Repair), then the rest by score.
A real problem with this list, worth knowing before you dial. Most of these are foundation-repair companies that also do leveling — only three are leveling-primary by name (Crown House, El Paso Concrete Leveling, Level Home). Under a CL-only year-one focus, that's a mismatch. It doesn't block this week — foundation guys with leveling crews are still real prospects and 17 accounts is enough for five days. But the list needs a proper CL-primary rebuild before week two. See the build queue.
The CRM's cold spine is not usable for dialling. sp_sales_contacts holds 175 cold accounts across A/B/C tiers, but the import kept only the last four digits of each phone number. Nothing has been touched since Jul-20. Fixing this is a build item, not a this-week item.
8 Scoreboard and kill rules
Score the funnel per stage. One blended conversion rate hides where the leak actually is.
| Stage | Metric | If it's weak, look at |
|---|---|---|
| Activity | Attempts logged | Calendar protection, avoidance |
| Reach | Connect rate | Phone accuracy, number reputation, time of day |
| Attention | Permission granted | Opener, pace, the first pause |
| Relevance | Right-fit conversations | List quality, niche language, hook quality |
| Commitment | Demo permissions + booked calls | Problem framing, offer clarity, the ask |
| Attendance | Held / booked | Reminders, time-to-meeting, demo delivery |
| Revenue | Signed clients | Diagnosis depth, proof, terms |
Auto-stop, no debate:
- Connect rate falls below ~5% → stop and investigate list and number health before touching a single word of the script.
- Zero positive replies after 50 touches → change the opener or the list. Not the volume.
- Any unusual cluster of STOPs or complaints → pause the channel, review the list source.
- Email bounce rate above 3% (once email is live) → pause sending, fix verification and auth.
Honest zeros. There is no outbound touch log yet — sp_touches is inbound attribution, not outbound activity. Until one exists, every dashboard will read zero for touches. That's the dashboard being honest, not broken. Log dispositions in GHL as you go; the 12 disposition types are already defined in sp_call_dispositions.
The outcome that matters this week is not dials. It's booked calls, and after that, signed clients. Compliments and "interesting, send me info" are not results.
9 What's blocked, and the build queue
Genuinely blocked right now:
| What | Why | When it clears |
|---|---|---|
| Cold email | DKIM and DMARC are both missing on stpierre.ai — verified in DNS today. Only SPF is published. No look-alike sending domain is registered. | ~3 weeks after you buy a sending domain and start warmup |
| Automated SMS sequences | Manual mode — auto_send FALSE, scripted SMS disarmed in production since Jul-27 | Your call. Recommendation: leave it off. This week's volume doesn't need it. |
| JV outreach (HMI, Koerner, Tyler Link, LevelRight) | Needs a published proof pack first. Pitching a partnership with one booked meeting is weak. | Once Dallas produces a real case-study number |
A2P is not a blocker and has not been since Jul-27. Both SMS numbers are operational, and the 626 number was proven delivering to your phone on Jul-28. Any doc still listing "A2P / carrier registration" as a gate is stale — including the Jul-28 CL market map, which I'll correct. Don't let it resurface.
Build queue, in the order it should happen:
- 1Buy and warm a sending domain. A look-alike like
stpierremarketing.comorgetstpierre.com— never the primary. Publish SPF + a 2048-bit DKIM key + DMARC atp=none. Start warmup at ~5 sends/day. This has a 2–4 week clock on it, so starting it this week is what makes mid-August possible. - 2Build the outbound touch log. One table, one row per dial, voicemail, text, and reply. Without it the scoreboard cannot tell you anything and every "we tried that" argument is unwinnable.
- 3Rebuild the list CL-primary. Google Maps scrape filtered on leveling / mudjacking / slabjacking / polyjacking, one state at a time, ~5,000 rows max per pass. Enrich mobile vs landline before any text touch. Small and verified beats big and generic every time.
- 4Backfill real phone numbers into
sp_sales_contactsfrom the 133-account tracker so the CRM's cold spine becomes dialable. - 5Rewrite the 10 Tier-A drafts to the partner frame. The $975/wk and $1-start copy in them is retired and would actively hurt you.
10 Sources
Built from a full pass over the brain: 326 matching video_notes, 85 documents, 43 session_handoffs, 38 deliverables, plus live Meta insights and live DNS.
Primary St. Pierre sources: the Cold-Calling Field Guide (Jul-14) — screenplay, territory gates, objection turns, dispositions, delivery coaching, all still canonical; the CL Market Map + JV Strategy (Jul-28); the Dream-900 Activation Pack and Week-1 call sheet (Jul-23); Sales Mechanics (discovery SOP, objection isolation, Hormozi closes); the SP Demo Bot README.
Corpus synthesized: the Glencoco cold-calling set (185-opener test, locked-room benchmark, Hormozi), the cold-email set (deliverability math, capacity math, reply-based CTAs), Josh Braun's Sales DNA plays, the cold-SMS GHL build, Conversionly lead-gen training.
Corrected in this pass:
- The $15/booked-lead claim — unsupported, replaced with $25/qualified lead.
- A2P listed as an infra landmine in the Jul-28 CL doc — resolved since Jul-27.
- Two call-sheet rows with refuted hooks still sitting in the dial order.
- The Week-1 list being described as CL when it is mostly foundation-repair.