Four Categories of Recommendations

After reviewing the current systems, codebase, analytics, and team structure — and meeting with both the executive team and the engineering team — I'm presenting four categories of recommendations:

1. Immediate fixes before March UPW — cart validation, portal access, reporting
2. Organizational restructuring — Build vs. Run teams
3. Growth strategy — from event-dependent to recurring revenue
4. Architecture modernization — Q2-Q4 roadmap

What's Already Fixed (That You May Not Know)

Before recommending changes, it's important to acknowledge work already completed by the engineering team:

Immediate Fixes Before March 12

2A. Checkout — Validate, Load Test, and Harden

Current state: The checkout is functional. Apple Pay is re-enabled, European customers can purchase, and the GTM bottleneck is resolved. The immediate fires are out.

Remaining risk: The checkout uses a multi-layered iframe architecture that is inherently fragile. When a customer clicks "Order Now," it loads an iframe from tonyrobbins.com/widget/order-form, which then loads 5+ additional Stripe iframes inside it (address fields, payment fields, Google Pay, hCaptcha, Google Maps autocomplete). That's up to 4 levels of iframe nesting across 6+ domains. We verified it works today — but it broke repeatedly under load during the summit. The architecture is vulnerable to browser security updates, ad blocker changes, third-party cookie deprecation, and cross-origin policy shifts.

What we need before March:

1. Comprehensive load testing from real browsers worldwide — not server-side stress tests. Actual browser-based synthetic testing:
   • 30,000+ concurrent sessions from US, UK, Germany, Australia, Brazil
   • Chrome, Safari, Firefox, Brave, iOS Safari, Android Chrome
   • With common ad blockers and privacy extensions enabled
   • Measuring page load, checkout render time, and payment completion
2. Pre-event checkout validation checklist: Apple Pay e2e, Google Pay e2e, credit card from US/UK/EU, mobile browsers, desktop with ad blockers, quantity > 1, promo codes, error states
3. Continuous synthetic monitoring: Automated checks every 5 minutes against live checkout. Alert Run team if load exceeds 3 seconds or checkout fails.

Recommended tools: Checkly or Catchpoint for synthetic monitoring. k6 with browser module or BrowserStack for real-device load simulation. These become permanent infrastructure.

Longer-term (Q2): Replace the iframe-based checkout with Stripe Payment Element — a direct integration not locked to a domain whitelist, removing the 4-layer iframe nesting entirely.

2B. Portal Access — From 3 Minutes to 20 Seconds

Problem: Customers wait 3-5 minutes (previously hours) for confirmation email and portal access. During a live event selling the next event, even 3 minutes feels broken.

Root cause:

1. Order enters Heroku job queue (~1.5 jobs/second at peak)
2. Email triggered through HubSpot's API
3. HubSpot rate limit: 200 calls per 10 seconds
4. At summit scale, HubSpot becomes the bottleneck

Solution: Send the confirmation email directly from the order processing API (SendGrid, Postmark, or AWS SES) and log the email-sent event back to HubSpot as an activity on the contact record. Customer gets email in ~20 seconds. Marketing still sees the event in HubSpot. No data is lost.

Prerequisite: IP warm-up for the new sending domain. 30 days is sufficient — this needs to start immediately.

2C. Landing Page — Dual-Deploy Strategy

Problem: Marketing builds pages in HubSpot for flexibility. But HubSpot pages are slow — heavy CMS overhead, third-party scripts, no edge caching.

What we already proved: I rebuilt the UPW landing page as a static page deployed to Cloudflare Pages — sub-second load, served from 300+ edge locations worldwide, identical design.

Proposed workflow:

1. Marketing continues building pages in HubSpot (no disruption)
2. Before major events, "snapshot" the final page and deploy to Cloudflare Pages
3. DNS points traffic to Cloudflare during high-traffic events
4. After event, revert to HubSpot for ongoing edits

Marketing keeps their tools. Customers get a fast page. Engineering gets peace of mind.

2D. Salesforce Reporting — Use What You Already Have

Problem: Leadership can't see real-time sales data during events. Salesforce reports are 3-4 hours behind.

Solution: Build 2-3 QuickSight dashboards against the operational database (live Heroku Postgres) for real-time event reporting. Key dashboards: live sales by SKU/time, conversion funnel, revenue by geography.

Organizational Restructuring — Build vs. Run

The Core Problem

The current team structure combines everyone into one group that does everything: operations, maintenance, new development, event support, DevOps, QA, and firefighting. The result:

• Developers get pulled to events to set up kiosks
• The same people running production are building new features
• No sprint has been closed in 3 weeks
• The team worked 3-4 weeks straight of 10-hour days and burned out
• Half the team is currently sick or on vacation
• When fires happen, all development stops
• When development is prioritized, operational issues pile up

This is not a people problem. This is a structure problem. The current structure guarantees failure regardless of who fills the roles.

Current Structure

Proposed Structure: Build Team + Run Team

How It Works in Practice

Key Principles

1. Run team never builds new features. They fix, maintain, monitor, and support.
2. Build team never gets pulled for event ops. They plan, build, test, and ship.
3. Event operations use contractors, not developers. Kiosks are not engineering work.
4. QA is consolidated — one supports Run, one supports Build.
5. Spork leads the Run team and stays as architecture authority across both teams.
6. Justin leads the Build team and stays as product authority — roadmap, priorities, UX decisions.

Budget Impact

3D. AI Engineering Upskill — The $20K Force Multiplier

Before hiring anyone, invest in the team you already have. A 2-day intensive AI Engineering Flow workshop led by Cole Medin — one of the leading practitioners in AI-native development — will bring every engineer up to a new standard of building with AI.

What this covers:

• AI-assisted development workflows — how to write, test, and ship 3-5x faster using AI coding tools
• Building AI agents — how to create autonomous agents that handle repetitive engineering and operational tasks
• Prompt engineering for production systems — not chatbots, but real AI integrated into business workflows
• Evaluating build vs. buy — when to build custom agents vs. use existing tools

The goal isn't to replace engineers — it's to turn each one into a team of five. A $20K investment that potentially saves $300K+ in hiring.

3E. AI Agents — Augmenting the Team Before Expanding It

After the workshop, the next step is an AI agent audit of the current roadmap and operational workflows. Before filling any open headcount, we inspect every planned hire and ask: can an AI agent handle 80% of this role?

The process:

1. Map current workflows — document what each team does daily, weekly, monthly
2. Identify agent opportunities — repetitive tasks, data processing, monitoring, reporting, triage
3. Build and deploy agents — starting with highest-impact, lowest-risk automations
4. Evaluate remaining gaps — only hire for what agents genuinely can't do

Likely agent candidates:

• Monitoring Agent — watches production systems, triages alerts, escalates only real issues (reduces DevOps load)
• Integration Agent — handles Salesforce sync, data pipeline failures, retry logic (reduces integration engineering load)
• QA Agent — automated regression testing, visual diff checks, pre-deploy validation
• Reporting Agent — generates daily/weekly dashboards, pulls data from lake, formats for leadership
• Support Agent — handles L1 internal support tickets, routes to right person, maintains knowledge base

Growth Strategy — From Event-Dependent to Recurring Revenue

The Structural Problem

The business model is almost entirely event-dependent and transactional. Revenue depends on filling arenas and virtual rooms, event by event, campaign by campaign.

When the checkout breaks at the summit, revenue falls off a cliff — because there's no recurring base to cushion it. The 10% YoY decline isn't just a marketing problem. It's a model problem.

The existing subscriptions (Inner Circle $97, Tony AI $99, Business Accelerator $83/mo) are low-priced, loosely connected to the event experience, and not positioned as the natural next step. They feel like add-ons, not a path.

4A. The "Mastery Path" — A Unified Progression System

The gap: Someone attends UPW, gets fired up, goes home, and... nothing structured happens. There's no guided bridge between the $495 event high and ongoing transformation.

The concept: A single, structured progression system that every customer enters at their first event.

• Level 1: Foundation — Post-event implementation program + Tony AI + community. Included 90 days post-event, then $149/mo.
• Level 2: Growth — Advanced content, peer accountability groups, monthly group coaching. $249/mo.
• Level 3: Mastery — Bridge to Results Coaching or Business Accelerator. $497/mo.
• Level 4: Leadership — Teach, lead communities, become certified. Application-based.

4B. Post-Event Implementation Program — The Missing Bridge

The biggest drop-off happens in the 30 days after an event. The emotional peak fades, patterns return, people drift.

The product: A structured 90-day program starting the day the event ends. Not content. Not a replay. A daily practice system.

• Week 1-2: Integration — daily 10-minute exercises. Tony AI guides with personalized prompts.
• Week 3-8: Practice — weekly challenges, peer accountability pods (5-7 people), live group coaching every 2 weeks.
• Week 9-12: Momentum — 12-month goals, daily rituals, "transformation story" presentation. Celebration event.

Pricing: Bundled into Mastery Path ($149/mo) or standalone $397 one-time.

4C. Tony AI as Connective Tissue — Not a Standalone Product

Currently Tony AI competes with Inner Circle ($97/mo) and Business Accelerator ($83/mo) for the same wallet. Three low-cost subscriptions that feel similar from outside.

Reposition: Tony AI becomes the platform layer connecting everything:

• After UPW: Implementation coach with personalized daily prompts
• Before next event: Reviews what you learned, sets intentions, identifies breakthrough areas
• During coaching: Reinforces sessions between calls, tracks commitments
• In the portal: Intelligent layer on all content — recommends based on history and progress
• For sales: Qualifies leads via conversation before a human calls

Tony AI stops being a $99/mo product and becomes included in every Mastery Path tier. Its data becomes the most valuable asset in the business.

4D. Annual Event Passport — Predictable Revenue

Every virtual event is sold individually, requiring a full acquisition campaign each time.

• Event Passport — All virtual events for one year: $1,997/yr
• Event Passport + Mastery Path — All events + progression: $2,997/yr
• Event Passport VIP — All virtual + one in-person event: $7,997/yr

4E. Load Testing as Revenue Protection

The summit cost millions because no one knew it was broken until customers reported it. Implement permanent synthetic monitoring and load testing: 30K+ concurrent sessions from 12+ countries, real browsers, real devices. Pre-event confidence reports to leadership.

4F. Speed-to-Lead: 8 Minutes → 60 Seconds

Lead routing flows through HubSpot → Salesforce → LeanData → sales rep. 7-8 minute delay. Responding in 1 minute vs. 5 increases conversion 3-5x.

Fix: Bypass Salesforce for routing. Webhook from HubSpot to a lightweight service that notifies the rep via SMS/Slack in 30 seconds. Better yet: Tony AI as first responder — qualifies the lead and warms them before the call.

4G. Activate the 700K Summit Leads + 4.5M Dormant Database

700K new leads (now): "Bring a friend" referral, SMS sequences, Meta/Google retargeting, physical mail to 33K cart abandoners.

4.5M dormant (Q2): Segment by value (phone call vs. email vs. direct mail), use the data lake, run a reactivation micro-event, build lookalike audiences.

4H. European Market + Local Payment Methods

Enable Klarna (installments), SEPA Direct Debit, iDEAL, Bancontact in Stripe. 1-2 days of configuration. Target EU summit attendees who couldn't checkout before.

4I. Real-Time Event Intelligence

Real-time dashboards during events enable mid-event offer adjustments, flash promotions to segments, and data-driven decisions instead of panic. Half of the panic comes from not knowing.

4J. Corporate/B2B Channel

No product exists for teams. A B2B subscription: team Tony AI access, quarterly workshops, custom Mastery Path tracks, engagement dashboards. $500-2,000/seat/year. One 50-person company = $50K replacing 100 individual subscriptions.

Architecture Roadmap — Q2-Q4 2026

The Tony Robbins OS — The Brain of the Business

Every system RRI runs today — Salesforce, HubSpot, Heroku, the portal, coaching tools, event platforms — is a disconnected island. Data lives in silos. Teams context-switch between 10+ tools. No single view of a customer exists.

The vision: Build a Tony Robbins Operating System (TROS) — a centralized intelligence layer that connects every system, powers every team, and gives every employee a custom work interface built for their role.

Backstage — Role-Based Dashboards:

• Executive Dashboard — real-time revenue, event sales, MRR/ARR, coaching pipeline, agent activity, escalations
• Events Dashboard — logistics checklists, attendee management, real-time sales during events, satisfaction tracking
• Sales Dashboard — lead pipeline, speed-to-lead metrics, conversion rates, Tony AI qualification handoffs
• Marketing Dashboard — campaign performance, audience segments, page speed, funnel analytics
• Coaching Dashboard — member progress, session notes, health scores, renewal predictions
• Operations Dashboard — system health, integration status, ticket queue, SLA tracking

Frontstage — Customer-Facing Intelligence:

• Tony AI Concierge — the unified customer touchpoint across web, app, SMS. "What program is right for me?" / "What's my Mastery Path progress?" / "Connect me with someone in my cohort"
• Unified Portal — single sign-on across all products. Progress tracking, event history, content recommendations powered by the shared brain

The Shared Brain — a single data layer powering everything:

• Customer profiles (enriched continuously from every touchpoint)
• Event history and transformation metrics
• Coaching program content and progress
• Sales and marketing intelligence
• Operational knowledge base (playbooks, runbooks, escalation procedures)

Prerequisites — Planning Period:

1. Executive requirements sessions — 2-3 workshops with leadership to define what each department needs from the OS
2. Justin leads product roadmap — translates requirements into a phased build plan with the Build team
3. Data architecture review — ensure the data lake can serve as the shared brain
4. Agent inventory — map which AI agents are needed per department (output of the AI Engineering Workshop)

Q2 2026 (April-June)

1. Heroku Migration — Contract ends September. Move order ingestion API and job queue to own virtualization (Sacramento/San Diego). Already planned.

2. Checkout V3 — Replace custom Stripe cart entirely. Stripe Payment Element with native Apple/Google Pay. Works from any CMS. Mobile-first, i18n for Europe.

3. Real-Time Data Pipeline — Increase lake sync to near-real-time. Build executive dashboards in QuickSight.

Q3 2026 (July-September)

4. Portal Unification — Merge 11 customer portals into single Tony Robbins Experience. New SSO. Built on Sanity + Next.js.

5. Salesforce Rationalization — Move reporting to data lake/QuickSight. Salesforce becomes CRM + finance only. Evaluate Blackthorn replacement with batch optimization (200-record bulk loads).

Q4 2026 (October-December)

6. Marketing Independence — Evaluate dedicated page builder (Webflow, Unbounce, or enhanced Sanity). Amplitude integrated across all touchpoints. Audience building and export.

7. CRM & Sales Engagement Review — Evaluate Salesloft/Outreach to replace Salesforce Sales Engagement. Target: lead response under 60 seconds.

Process & Planning Discipline

Quarterly Technology Planning

1. What we accomplished — demo, not slides
2. What's on the roadmap — with effort estimates and dependencies
3. What we need — budget, hiring, decisions from leadership
4. What we're saying no to — and why

Sprint Discipline

• Build team: 2-week sprints. Every sprint has a goal. Every sprint closes.
• Sprint demos every 2 weeks — leadership can attend
• Mid-sprint requests go to backlog unless P1
• Run team handles all urgent mid-sprint requests

Communication Cadence

Change Management for the Executive Team

The hardest change isn't technical — it's behavioral:

• Stop making last-minute decisions that force fire drills. 30-day freeze before major events.
• Commit to KPIs and don't change them every 3 weeks. The data lake is useless if what you measure keeps shifting.
• Trust the process. If Sprint 4 says Feature X ships, it ships in Sprint 4 — not tomorrow because someone panicked.
• Budget for run, not just build. Operations costs money. That's not waste — it's the cost of keeping lights on while you build the future.

What I Recommend We Do Together

Phase 1: March UPW (Now → March 12)

• Set up load testing + run full pre-event simulation (30K+ sessions, global, real browsers)
• Implement continuous synthetic monitoring on all customer-facing pages
• Validate checkout end-to-end: Apple Pay, Google Pay, EU, mobile, ad blockers
• Implement direct email sending (bypass HubSpot for transactional emails)
• Deploy optimized landing page via Cloudflare for UPW traffic
• Build 2-3 real-time event dashboards
• Enable Klarna and European payment methods in Stripe
• Publish pre-event confidence report to leadership

Phase 2: Restructuring + AI Upskill (March → April)

• Present Build vs. Run recommendation to executive team
• Map existing team members to new structure (Spork → Run, Justin → Build)
• Run AI Engineering Flow workshop with Cole Medin ($20K) — upskill entire engineering team
• Conduct AI agent audit — inspect current roadmap for agent opportunities
• Evaluate remaining hiring needs after agent deployment plan
• Establish quarterly planning cadence
• First quarterly roadmap session

Phase 3: Tony Robbins OS + Growth (Q2-Q4)

• Executive requirements workshops — define what each department needs from the OS
• Justin leads product roadmap for TROS — phased build plan
• Map current org chart → create AI Org Chart vision → plan AAA phases
• Build and deploy first AI agents (monitoring, integration, QA, reporting)
• Guide Salesforce rationalization and Heroku migration
• Design and launch Mastery Path + Event Passport products
• Reposition Tony AI as connective tissue across all products
• Begin TROS backstage dashboards — executive and events interfaces first
• Advisory role: monthly check-ins, architecture reviews, vendor evaluations

Landing Page Proof of Concept

We already rebuilt the UPW landing page as a demonstration:

• Original (HubSpot): go.tonyrobbins.com/march-upw — slow, render-blocking scripts, heavy DOM
• Optimized (Cloudflare): upw-landing.pages.dev — sub-second load, static, edge-cached globally

The Cloudflare version is an exact visual replica with all CSS inlined, fonts preloaded, images served from CDN, no render-blocking JavaScript, and served from 300+ edge locations.

Known limitation: The Stripe checkout widget has a CSP frame-ancestors whitelist that only allows embedding on specific domains: tonyrobbins.com, go.tonyrobbins.com, *.hubspot.com, *.hs-sites.com, *.hubspotpagebuilder.com, and timetorisesummit.com. Any non-whitelisted domain (including our Cloudflare Pages deploy) is blocked — the checkout drawer opens but the iframe refuses to load. This proves the architectural fragility: the checkout is domain-locked, requiring either whitelist addition or a rebuild as a direct Stripe Payment Element (recommended in Part 2A).

This dual-deploy approach can be applied to any sales page before major events.