Project WE: Probability-Based Messaging System & Cross-Pollinator Development Review

Date: March 18, 2026

Participants: Larry Czerwonka, Jeffrey Rudnick, Jon, Stacey, Melissa Krogman, Auto Doctor Front Desk (Joe)

TASK COMMITMENTS BY PERSON

Larry Czerwonka:

• Will send links to both demo pages for team testing and feedback (Timeline: after call)

(Larry Czerwonka): "I'll send links. Please go play with the 2 pages that we'll send. See what we're missing. See if anything comes to mind."

• Will send mascot psychology write-up explaining why effective mascots must be non-human and memorable (Timeline: after call)

(Larry Czerwonka): "I'll write up a two-page document on it and then send it to everyone so everyone can understand that... That's the key is you're trying to build, you're trying to create something that people see and go, I know who that is. So they pay attention."

• Will convert Auto Doctor's preferred mascot to 3D if full-body source artwork exists — charges for recreation work if only partial artwork available (Timeline: pending Joe confirmation of artwork availability)

(Larry Czerwonka): "But yeah I can convert it. The question is, if you send me just a head shot, then I got to invent the body. If you can find something that is actually a full body version of the character, I can convert that quicker."

• Will add "write your own offer" text input field at top of cross-pollinator onboarding form above "not sure" button — addresses direct-input gap identified during demo (Timeline: unspecified, low complexity)

(Larry Czerwonka): "So I might as well just add at the top, just a text box, and it says, you know, what's the offer you want to make? And then put that in and then have the text, not sure... so we'll have a text box... I can do that."

• Will fix URL display truncation in cross-pollinator onboarding — currently showing full encoded URL instead of clean display (Timeline: unspecified, minor fix)

(Larry Czerwonka): "The URL is there... I'm supposed to be peeling that out. It should not look like all that crap... I'm supposed to just show DoorDash or whatever they what the name is and then it shows the name on top of it."

• Will fix image auto-scaling issues where developers keep re-adding width/height constraints he removes — creating repeating bug pattern (Timeline: unspecified, ongoing issue)

(Larry Czerwonka): "My developers keep adding width and heights and I keep saying you can't do that... they keep, they get assigned different jobs... This is a thing that keeps happening."

• Will implement messaging interface editing capability and reporting dashboard — committed to live deployment by next week, March 25, 2026 (Timeline: 1 week)

(Larry Czerwonka): "That's what next week is about... by the end of next week, they should be able to change all of this... Next week there'll be editing."

• Will complete cross-pollinator onboarding flow for shops to send businesses to self-service signup (Timeline: end of March 2026)

(Larry Czerwonka): "By the end of the month... you should be able to send them to a page that does all of this, right? And then it would then trigger it to send an email to Melissa with all the information."

• Will dedicate 2 full days per week to messaging system development — accelerating deployment timeline (Timeline: ongoing commitment)

(Larry Czerwonka): "I'm spending 2 full days a week on just the messages... I'm doing two days a week of me personally just on this."

• Will evangelize Arivu Protocol transcript system improvements to internal team — promoting public/internal two-copy format and enhanced documentation approach (Timeline: ongoing, evangelism activity)

(Larry Czerwonka): "It's a two file system. It's not a different thing. It's I made a copy... So anyway, that's in my notes, but that came out really good... So it came out good. I'm excited about how that came out."

Jeffrey Rudnick:

• Will test both demo interfaces (messaging system and cross-pollinator onboarding) and provide feedback on missing elements or UX improvements (Timeline: after receiving links)

(Larry Czerwonka): "Please go play with the 2 pages that we'll send. See what we're missing. See if anything comes to mind."

Jon :

• Will test both demo interfaces and provide shop owner perspective on workflow and missing features (Timeline: after receiving links)

(Larry Czerwonka): "Please go play with the 2 pages that we'll send. See what we're missing. See if anything comes to mind."

• Waiting for cross-pollinator tool completion before recruiting DeKalb County fundraiser partners — confirmed tool solves actual friction blocking adoption (Timeline: end of March 2026 deployment target)

(Jon): "The event is like later on in April and I need this tool before I can start reaching out to people... At least I'd have it ready for them to start entering their information."

Melissa Krogman :

• Will test cross-pollinator onboarding interface from graphics vendor perspective and confirm workflow meets business requirements (Timeline: after receiving links)

(Larry Czerwonka): "Please go play with the 2 pages that we'll send. See what we're missing. See if anything comes to mind."

Joe:

• Will ask Auto Doctor owner Dustin to provide full-body source artwork for preferred mascot character — needed for Larry's 3D conversion (Timeline: unspecified)

(Larry Czerwonka): "Joe, have Dustin actually give us the actual image that he wants to turn into a mascot and see if it's, if the person is like in a full body or if it's cropped at the waist or whatever."

NEXT 72 HOURS - CRITICAL PATH

CRITICAL

Larry: Deploy Demo Page Links & Mascot Psychology Write-Up

DELIVERABLE: Send both demo page links (probability messaging interface + cross-pollinator onboarding) to team for testing plus mascot psychology explanation document.

PURPOSE: Enable team feedback collection before next development sprint, educate all shops on effective mascot design to prevent repeat of Auto Doctor debate.

TIMELINE: Immediate post-call.

LINKS:

HIGH PRIORITY

Larry: Complete Messaging Interface Editing + Reporting Dashboard

TARGET: Live (prototype) deployment by March 25, 2026 (next week).

WORK REQUIRED: Implement shop owner ability to edit all 35+ message templates, build reporting dashboard showing message performance by customer grade, deploy to Project WE shops first for validation.

RESOURCE ALLOCATION: 2 full days per week dedicated to messaging system development.

CONSEQUENCE OF DELAY: Blocks Project WE validation before 200+ shop rollout, pushes competitive moat feature deployment timeline, damages credibility with core shop partners testing untouchable features.

Larry: Complete Cross-Pollinator Onboarding Self-Service Flow

TARGET: End of March 2026 deployment.

WORK REQUIRED: Add direct offer input field, fix URL display truncation, resolve image scaling bug pattern, complete graphics request auto-email to Melissa, deploy self-service signup page shops can send partners to.

EXTERNAL DEPENDENCY: Jon's DeKalb County fundraiser (April event) blocked waiting for this tool — can't recruit partners until system exists to onboard them without manual work.

CONSEQUENCE OF DELAY: Jon misses recruitment window for April fundraiser partners, proves friction still exists blocking cross-pollinator adoption at scale.

Joe: Obtain Auto Doctor Full-Body Mascot Source Artwork

TARGET: Ask Dustin to provide complete source artwork for preferred mascot character.

PURPOSE: Enable Larry's 3D conversion work — full-body artwork allows quick conversion, partial/headshot only requires Larry to invent body (billable recreation work).

BUSINESS IMPACT: Custom mascot 3D rendering positioned as billable work Auto Doctor will accept since it's outside standard platform offering.

All Team Members: Test Demo Interfaces & Provide Feedback

WHO: Jeffrey, Jon, Melissa, Joe testing both probability messaging system and cross-pollinator onboarding interfaces.

FOCUS: Identify missing features, UX confusion points, workflow gaps before next development sprint.

VALUE: Team catches issues during beta testing with core partners rather than discovering them during 200-shop rollout where damage multiplies.

Current Issues Identified During Demo:

MINOR UX GAP

Missing Direct Offer Input Path

Onboarding flow lacks obvious "I know my offer, let me type it" option at top. Currently forces everyone through suggested-offer-first route even when unnecessary. Creates UX confusion for partners who arrive knowing exactly what they want to offer.

(Jon): "If I know exactly what I want to offer... I would just be confused by not sure."

Larry's Fix: Will add text input field at top of form labeled "What's the offer you want to make?" with "Not sure" button below as fallback option. Simple addition, low complexity.

DISPLAY BUG

URL Display Truncation

System showing full encoded URL instead of clean business name display. Should show "DoorDash" or business name cleanly, not entire URL string with parameters.

(Larry Czerwonka): "The URL is there... I'm supposed to be peeling that out. It should not look like all that crap."

Impact: Pure cosmetic issue, doesn't block functionality but looks unfinished/unprofessional in demo.

RECURRING ISSUE

Image Auto-Scaling Keeps Breaking

Developers repeatedly re-add width/height constraints Larry removes, creating repeating bug where images don't scale properly. Suggests code review or testing gaps in development process.

(Larry Czerwonka): "My developers keep adding width and heights and I keep saying you can't do that... This is a thing that keeps happening."

Root Cause: Developers getting assigned different jobs touching same code without clear documentation or testing protocol. Larry fixes issue, different developer touches code later and re-introduces constraint. Pattern indicates process problem, not individual competence issue.

BUTTON LABEL CLARITY

"Not Sure" Button Ambiguity

Jeff questioned whether "Not sure" button generates offer suggestion or is literal "I don't know" response. Needed Larry to explain it generates suggested offer which then appears with "Change offer" option.

(Jeffrey Rudnick): "The not sure button that you hit there, is it creating that offer suggestion or is that a pre written template?"

Larry's Explanation: "Not sure" button generates suggested offer automatically which then appears with "Change offer" option allowing full customization. So direct input exists, just accessed through suggested-offer-first path. Design choice prioritizes reducing blank-page paralysis by showing example first.

Friction Factors

PROCESS GAP

Developer Discipline Issues with Image Scaling

Image scaling keeps breaking because developers re-add constraints Larry removes. Indicates code review/testing gaps where changes aren't validated before deployment. Pattern suggests lack of clear documentation or communication about why constraints were removed.

(Larry Czerwonka): "My developers keep adding width and heights and I keep saying you can't do that... This is a thing that keeps happening."

Impact: Creates rework loops and erodes Larry's trust in developers. "This is a thing that keeps happening" signals frustration with recurring process failure. Not critical path blocker but indicates broader quality control weakness.

CLIENT EDUCATION GAP

Mascot Psychology Resistance

Auto Doctor owner rejecting designed mascot for generic alternative despite Larry's memorability arguments. Represents broader client education challenge: shops prefer familiar/comfortable over memorable/effective because they don't understand psychology of pattern interruption in advertising.

(Larry Czerwonka): "I get that people don't get it, but people don't understand the psychology of people."

Impact: Client makes suboptimal branding choice that will reduce marketing effectiveness. Larry's response notable: pushed back with education, agreed to execute for billable fee, sending psychology write-up to entire team as teachable moment preventing repeat with other shops.

UX POLISH

Missing Direct Offer Input Path

Onboarding flow lacks obvious "I know my offer, let me type it" path, forcing everyone through suggested-offer-first route. Creates UX confusion for partners who arrive with predetermined offer.

Impact: Minor friction, easily fixed with text input field addition. Jon's feedback valuable: "If I know exactly what I want to offer... I would just be confused by not sure." Simple UX improvement preventing user confusion.

EXTERNAL PRESSURE

Timeline Dependency Chain - Jon's DeKalb County Fundraiser

Jon's DeKalb County fundraiser commitment (April event) blocked waiting for cross-pollinator completion. External commitment creates pressure on development schedule — can't slip timeline without impacting Jon's ability to deliver on fundraiser promise.

(Jon): "The event is like later on in April and I need this tool before I can start reaching out to people."

Impact: Positive pressure (proves real demand) but creates hard deadline Larry must hit. Missing end-of-March target means Jon can't recruit partners in time for April event, damages credibility with Project WE shop owner serving as living laboratory.

KEY DISCUSSION POINTS

Probability-Based Messaging System Architecture

Larry demonstrated visual workflow interface showing 35+ automated messages organized by customer grade (VIP, Great, Loyal, Good, New, Slipping, At-Risk, Inactive) and triggered by return probability calculations rather than arbitrary time intervals. System fundamentally different from traditional loyalty program blasts because it forces differentiated treatment based on customer value.

Core Philosophy - "Stop Treating Everyone the Same":

Value Concentration Reality: 50% of shop income comes from 16% of customers, yet most shops spend marketing budget proportionally to customer count (treating D/F-grade customers consuming 80% of budget while generating <20% revenue)
Forced Differentiation: Messages are grade-specific by architectural requirement — can't send everyone same blast anymore because system won't allow it, must explicitly define what VIP gets vs what one-and-done gets
Probability-Based Triggering: System sends messages when customer is 70% likely to return (calculated from visit history patterns) instead of "send blast every 90 days" arbitrary timing that ignores individual customer behavior
Intelligence Accumulation: Platform grades which communications actually drive return visits over time, learning what works per customer segment — knowledge that resets to zero if shop switches platforms

(Larry Czerwonka): "50% of your income is from 16% of your customers. You need to quit treating everybody the same and only treat, you know, take care of those people the way they deserve to be taken care of."

Message Categories and Triggers:

VIP/Great Customers (A/B grades): Premium offers, birthday recognition, exclusive early access, personal thank-you messages — high-touch communications justifying high spend
Loyal/Good Customers (C/D grades): Standard rewards, maintenance reminders, seasonal promotions — moderate touch maintaining relationship
New Customers: Welcome series, education content, second-visit incentives — converting one-time visitors to repeat customers
Slipping/At-Risk (declining visit frequency): Re-engagement offers, "we miss you" outreach, win-back campaigns — targeted intervention before they become inactive
Inactive: Minimal automated low-cost attempts, no premium offers or personal outreach — acknowledging sunk cost and moving resources to viable customers

Current Development Status:

Visual workflow interface complete showing all 35+ message paths
Backend probability calculations operational and triggering messages
Editing capability deployment target: next week (March 25, 2026)
Reporting dashboard showing message performance by grade: next week
Larry dedicating 2 full days per week to accelerate completion

Cross-Pollinator Onboarding Automation

Web-based partner business signup collecting offer details, pulling 6 images from business website + 3 stock options automatically, generating graphics request to Melissa's team. Solves major friction: shops currently can't recruit cross-pollinator partners until tool exists to handle their signup/onboarding without manual shop owner work.

Workflow Design:

Step 1 - Business Information: Partner enters business name, website URL, contact info
Step 2 - Automated Image Harvest: System pulls 6 images from website automatically, adds 3 stock images from library, displays 9 total options for partner selection
Step 3 - Offer Definition: Two paths: (a) "Not sure" button generates suggested offer which appears with "Change offer" option allowing full customization, or (b) direct text input field for partners who know exactly what they want
Step 4 - Graphics Request Generation: Completed form auto-emails Melissa's team with all collected information, selected images, offer details — no manual handoff required

Image Harvesting Technical Details:

System fetches website metadata and scrapes for images automatically
Success rate approximately 60% currently, targeting 80% before launch
When image harvest fails, stock library provides fallback options
Partners can upload their own images if auto-fetch doesn't find suitable options
All images preview at actual size in selection interface

Mascot Psychology Debate

Auto Doctor owner (Dustin, via Joe) rejected Larry's designed mascot in favor of generic white male character someone else created. Larry pushed back on psychology/memorability grounds but ultimately agreed to 3D-render Dustin's choice for billable fee if source artwork available.

(Larry Czerwonka): "I get that people don't get it, but people don't understand the psychology of people... it's a white guy that looks like everyone else that is not memorable. A mascot needs to be memorable or it doesn't work. It's just another ad."

Larry's Mascot Effectiveness Principles:

Can't look like anything expected or it blends into background noise
Must be something that makes people go "that's different" so they pay attention
Examples that work: Michelin Man (not guy in tires), Pillsbury Doughboy (not big fat guy), Geico Gecko (not insurance person), Bachman's B (not mechanic)
Human mascots that worked: Marlboro Man only, but anyone under 40 doesn't recognize reference anymore
Goal is instant recognition: "You see it and you go, I know who that is. It's not Geico person."

(Larry Czerwonka): "That's the key is you're trying to build, you're trying to create something that people see and go, I know who that is. So they pay attention. That's that's all a mascot's about."

Resolution: Larry will send psychology write-up explaining mascot principles to entire team, will convert Dustin's preferred mascot to 3D if full-body source artwork exists (Joe to check), will charge for recreation work if only partial artwork available. Jeff positioned as billable work Auto Doctor will accept since it's custom request outside standard platform.

Development Momentum

STRONG (7.5/10)

Both systems reaching beta completion with aggressive end-of-March deployment timeline