# Customer Support Cost Analysis — RateRight
**Date:** 2026-02-26 | **Author:** Harper (Finance/Legal)
**Context:** Platform launch imminent. Zero support infrastructure. Need to model when support costs force hiring decisions.

---

## EXECUTIVE SUMMARY

**RateRight doesn't need human support staff for at least 6-12 months.** At projected volumes (30-190 active users by Month 1, scaling to 500-1,500 by Month 6), ticket volumes will be 10-75/month — well within what Michael + AI agents can handle for free.

The critical threshold is **~200 tickets/month** — that's when Michael's time cost exceeds outsourced support, and **~500 tickets/month** when dedicated automation becomes essential. At realistic growth, that's Month 9-12.

**The real risk isn't ticket volume — it's ticket TYPE.** Marketplace disputes (no-shows, bad matches, payment issues) require human judgement. AI can't resolve "this worker didn't show up, I want my $50 back" without a clear policy. Define the refund policy before launch. That eliminates 40%+ of potential escalations.

---

## 1. TICKET VOLUME PROJECTIONS

### Industry Benchmarks

| Source | Metric | Value |
|--------|--------|-------|
| MetricNet | Top-performing support organisations | 0.5 tickets/user/month |
| 8020 Consulting | SaaS target | <10% of active users monthly |
| General marketplace benchmark | Two-sided platforms (early stage) | 15-25% of active users monthly |
| Labour/services marketplace | Higher-touch (disputes, no-shows) | 20-35% of active users monthly |

**Why RateRight will be at the HIGH end of benchmarks:**
- Two-sided marketplace = support from BOTH contractors AND workers
- Financial transaction involved ($50/hire) = payment disputes
- Physical work outcomes (no-shows, quality issues) = high-emotion tickets
- New platform, no FAQ/help centre = more "how do I...?" questions
- Construction users = phone-first, less comfortable with self-service

**My estimate: 20-30% of active users will generate a support ticket per month in the early period (Month 1-3), declining to 10-15% as self-service improves.**

### Ticket Volume by Growth Scenario

Using user projections from the 30-day model and national expansion model:

| Month | Active Users (Realistic) | Ticket Rate | Est. Tickets/Month | Tickets/Week |
|-------|------------------------|-------------|--------------------:|-------------:|
| **1** | 40-75 | 25% | 10-19 | 3-5 |
| **2** | 75-130 | 22% | 17-29 | 4-7 |
| **3** | 130-200 | 20% | 26-40 | 7-10 |
| **4** | 200-280 | 18% | 36-50 | 9-13 |
| **5** | 280-370 | 16% | 45-59 | 11-15 |
| **6** | 370-500 | 15% | 56-75 | 14-19 |
| **9** | 700-1,000 | 12% | 84-120 | 21-30 |
| **12** | 1,200-1,800 | 10% | 120-180 | 30-45 |
| **18** | 2,500-4,000 | 8% | 200-320 | 50-80 |

Ticket rate declines over time as:
- FAQ/help centre gets built
- Common issues get automated (password resets, status checks)
- Users become more familiar with the platform
- AI chatbot handles tier-1 queries

---

## 2. TICKET TYPE BREAKDOWN (Predicted)

### Two-Sided Marketplace Ticket Categories

For a construction hiring marketplace, tickets will split roughly:

| Category | % of Tickets | Source | Complexity | AI-Resolvable? |
|----------|-------------|--------|-----------|---------------|
| **Account/Signup issues** | 15-20% | Both sides | Low | ✅ Yes (90%+) |
| **How-to / platform navigation** | 15-20% | Both sides | Low | ✅ Yes (85%+) |
| **Payment/billing queries** | 10-15% | Contractors | Medium | ⚠️ Partially (60%) |
| **Match quality complaints** | 10-15% | Contractors | Medium-High | ❌ No — needs judgement |
| **No-show / dispute** | 10-15% | Contractors | High | ❌ No — needs policy + human |
| **Worker availability/profile** | 8-12% | Workers | Low-Medium | ✅ Yes (80%) |
| **Refund requests** | 5-10% | Contractors | High | ❌ No — needs approval |
| **Bug reports** | 5-8% | Both sides | Medium | ⚠️ Triage only, Builder fixes |
| **Safety/compliance concerns** | 2-5% | Both sides | Critical | ❌ No — immediate escalation |
| **Other** | 5-10% | Both sides | Variable | ⚠️ Mixed |

### The Key Insight: Two Tiers

**Tier 1 (60-70% of tickets):** Account, how-to, basic payment queries, profile help. These are automatable with a good FAQ, help centre, and AI chatbot. Cost to resolve: near $0 once built.

**Tier 2 (30-40% of tickets):** Disputes, no-shows, refunds, match quality, safety. These require human judgement, policy decisions, and sometimes real-time phone calls. These are what break a founder's time and eventually force a hire.

---

## 3. COST PER TICKET — THREE MODELS

### Model A: Michael Handles Everything (Month 1-6)

**Assumptions:** Michael handles all Tier 2 tickets personally. AI agents + basic FAQ handle Tier 1. No additional cost — just Michael's time.

| Metric | Value |
|--------|-------|
| Michael's implicit hourly rate | $0 (founder time, no salary) |
| Actual opportunity cost per hour | ~$50-100 (building product, sales, strategy) |
| Average time per Tier 2 ticket | 15-30 minutes (phone call or email chain) |
| Average time per Tier 1 ticket | 2-5 minutes (FAQ link, quick reply) |
| Monthly Tier 2 tickets (Month 3, realistic) | 8-16 |
| Monthly time commitment (Month 3) | 2-8 hours |
| Opportunity cost (Month 3) | $100-$800/month |

**Verdict:** Workable for Months 1-6. Michael spends 2-8 hours/month on support. At his schedule (90-min evening window), that's 1-5 evenings per month consumed by support instead of building. Manageable but eroding.

### Model B: Outsourced Support (When Michael's Time Is Saturated)

Australian outsourced support rates (2026 benchmarks):

| Provider Type | Hourly Rate | Per-Ticket Cost (est.) | Min Commitment |
|---------------|------------|----------------------|----------------|
| **Offshore (Philippines/India)** | $8-15/hr | $4-8/ticket | 20-40 hrs/month |
| **Nearshore (Fiji, NZ)** | $18-30/hr | $9-15/ticket | 20 hrs/month |
| **Onshore (Australia)** | $40-65/hr | $20-33/ticket | Varies |
| **Specialist marketplace support (e.g. PartnerHero)** | $25-45/hr | $13-23/ticket | 40 hrs/month |

**For RateRight, offshore makes no sense** — construction workers and contractors expect Australian English, understand AU regulations, and won't tolerate scripted responses about their $50.

**Realistic option: AU-based part-time virtual assistant or freelancer.**

| Scenario | Monthly Tickets | Hours/Month | Cost (at $45/hr AU) | Cost/Ticket |
|----------|----------------|-------------|---------------------|-------------|
| Month 6 (realistic) | 56-75 | 15-25 hrs | $675-$1,125 | $12-$15 |
| Month 9 | 84-120 | 22-35 hrs | $990-$1,575 | $12-$13 |
| Month 12 | 120-180 | 30-50 hrs | $1,350-$2,250 | $11-$13 |

**This is cheaper than hiring full-time ($4K/month) but still a meaningful new cost line.**

### Model C: Full-Time Support Hire

| Metric | Value |
|--------|-------|
| Junior support agent salary (AU) | $50-55K/year ($4,200-$4,600/month) |
| + Super (11.5%) | $483-$529/month |
| + Equipment, software, training | ~$200/month amortised |
| **Total loaded cost** | **$4,883-$5,329/month** |
| Capacity | ~160 hrs/month = 320-640 tickets (at 15-30 min each) |
| Cost per ticket | $8-17 |
| **Break-even vs outsource** | ~250-350 tickets/month |

**A full-time hire makes financial sense only above ~300 tickets/month.** Below that, outsourcing or founder-time is cheaper.

---

## 4. SUPPORT COST vs REVENUE CURVE

What percentage of revenue does support consume at each stage?

| Month | Revenue (Realistic) | Support Model | Support Cost | % of Revenue |
|-------|--------------------|--------------|--------------:|-------------:|
| 1 | $195 | Michael (free) | $0 | 0% |
| 3 | $440 | Michael (free) | $0 | 0% |
| 6 | $1,368 | Michael (strained) | $0 (but 8-15 hrs opportunity cost) | 0% nominal |
| 9 | $2,100 | Outsource part-time | $990-$1,575 | 47-75% ⚠️ |
| 12 | $2,589 | Outsource or hire | $1,350-$2,250 | 52-87% ⚠️ |
| 18 | $4,500 | Full-time hire | $4,883-$5,329 | 109-118% ❌ |

**The brutal math:** At Month 12-18 realistic volumes, human support costs consume 50-100%+ of revenue. This is why automation isn't optional — it's existential.

### Industry Benchmark: Support as % of Revenue

| Company Stage | Healthy Range | RateRight Target |
|---------------|--------------|-----------------|
| Pre-revenue / early | 0% (founder handles) | ✅ Current state |
| $1K-$5K MRR | 10-20% | Target: <15% |
| $5K-$20K MRR | 8-15% | Target: <10% |
| $20K+ MRR | 5-10% | Target: <8% |

**To stay under 15% of revenue at Month 12 ($2,589 MRR), support costs must be under ~$388/month.** That's only achievable with heavy automation.

---

## 5. AUTOMATION STRATEGY — THE AI-FIRST APPROACH

### Why This Is RateRight's Unfair Advantage

We already have an 8-agent AI fleet. Other marketplaces hire support humans first, then try to automate later. We should automate first and hire humans only for what AI genuinely can't handle.

### Automation Tiers

| Tier | Ticket Types | Solution | Cost | Timeline |
|------|-------------|----------|------|----------|
| **Tier 0: Self-service** | How-to, account, FAQ | Help centre / knowledge base | $0 (Builder builds it) | Pre-launch |
| **Tier 1: AI chatbot** | Account issues, status queries, basic payment Qs | AI agent (existing fleet or dedicated bot) | ~$20-50/month (API costs) | Month 1-2 |
| **Tier 2: Automated workflows** | Refund processing, no-show flagging, dispute initiation | Rule-based automation in platform | $0 (Builder builds it) | Month 2-4 |
| **Tier 3: Human escalation** | Complex disputes, safety, angry contractor on phone | Michael → eventually part-time VA | $0 → $675+/month | Month 6-9 |

### Projected Automation Impact

| Month | Total Tickets | Tier 0+1 (Automated) | Tier 2 (Semi-auto) | Tier 3 (Human) | Human Tickets/Month |
|-------|--------------|---------------------|--------------------|-----------------|--------------------|
| 1 | 10-19 | 0% (nothing built) | 0% | 100% | 10-19 |
| 3 | 26-40 | 40% (basic FAQ live) | 10% | 50% | 13-20 |
| 6 | 56-75 | 55% (chatbot live) | 15% | 30% | 17-23 |
| 9 | 84-120 | 65% | 15% | 20% | 17-24 |
| 12 | 120-180 | 70% | 15% | 15% | 18-27 |
| 18 | 200-320 | 75% | 15% | 10% | 20-32 |

**With aggressive automation, human ticket volume stays at 17-32/month even as total tickets hit 200-320.** That's 4-8 hours of Michael's time — manageable indefinitely without hiring.

### Automation Cost vs Hiring Cost

| Month | Total Tickets | Human-Only Cost (outsource) | Automated + Residual Human Cost | Savings |
|-------|--------------|----------------------------|-------------------------------|---------|
| 6 | 56-75 | $675-$1,125 | $50 (AI) + $0 (Michael handles 17-23) = **$50** | $625-$1,075 |
| 9 | 84-120 | $990-$1,575 | $50 + $0 = **$50** | $940-$1,525 |
| 12 | 120-180 | $1,350-$2,250 | $50 + $0 = **$50** | $1,300-$2,200 |
| 18 | 200-320 | $2,400-$3,840 | $50 + $675 (PT VA for overflow) = **$725** | $1,675-$3,115 |

**Automation saves $1,300-$3,100/month by Month 12-18 vs outsourcing everything.** Over 12 months, that's $15K-$37K — potentially more than the MVP Ventures grant.

---

## 6. HIRING DECISION THRESHOLDS

### When to Hire (Clear Triggers)

| Trigger | Metric | Recommended Action | Monthly Cost Impact |
|---------|--------|--------------------|-------------------|
| **Michael spending >10 hrs/month on support** | Track from Day 1 | Start building Tier 1 automation urgently | $0 (Builder time) |
| **Michael spending >20 hrs/month on support** | ~Month 6-9 (without automation) | Hire part-time freelance VA (AU-based, 10 hrs/week) | +$450-$650/month |
| **>100 human-requiring tickets/month** | ~Month 12-18 (without automation); unlikely with automation | Outsource Tier 2-3 to specialist | +$1,200-$2,000/month |
| **>300 human-requiring tickets/month** | ~Month 18+ at scale (unlikely with automation) | Full-time support hire | +$4,883-$5,329/month |
| **ANY safety incident** | Any time | Immediate: define safety escalation protocol | $0 (policy, not people) |
| **Revenue >$5K MRR** | ~Month 12-15 (optimistic) | Can afford part-time support (~10% of revenue) | $500/month budget |

### The Decision Matrix

```
Human tickets/month:   <25        25-50       50-100      100-300     300+
                        │           │           │            │          │
Action:              Michael    Michael +    PT VA       Outsource   FT Hire
                     handles    AI chatbot   (10hrs/wk)  specialist
                        │           │           │            │          │
Monthly cost:          $0         $50        $450-650   $1,200-2K   $4,883+
                        │           │           │            │          │
Revenue needed:        $0         $0         $3K MRR    $8K MRR    $16K MRR
(to stay <15% of rev)
```

### What NOT to Do

1. **Don't hire a support person before $3K MRR.** It'll consume 100%+ of revenue and 5.7x the burn rate. Use AI + Michael's time.

2. **Don't outsource offshore.** Construction workers expect to talk to someone who understands Australian construction. A Philippines-based agent reading a script about "your job posting experience" will destroy trust faster than no support at all.

3. **Don't ignore support metrics.** Track from Day 1: ticket volume, type, resolution time, Michael's hours. If you don't measure it, you'll wake up one day spending 20 hours/week on support and not know when it happened.

---

## 7. PRE-LAUNCH ACTION ITEMS (Support Readiness)

These cost $0 and prevent 40%+ of early tickets:

| # | Action | Owner | Impact | Status |
|---|--------|-------|--------|--------|
| 1 | **Write a 10-question FAQ** | Herald (content) + Harper (payment/refund Qs) | Eliminates 15-20% of tickets | Not started |
| 2 | **Define refund policy** | Harper (draft) → Michael (approve) | Eliminates 5-10% of tickets + enables automation | Not started |
| 3 | **Set up a support email** | Michael | Single inbound channel, trackable | Not started |
| 4 | **Create a no-show procedure** | Harper (policy) + Builder (automation) | Highest-emotion ticket type, needs a clear path | Not started |
| 5 | **Add help text to key UI flows** | Builder | Reduces "how do I...?" tickets | Unknown |
| 6 | **Stripe receipt verification** | Harper/Builder | Ensure $50 charge generates compliant receipt with ABN | Not started |

**Total cost: $0. Total ticket prevention: 30-40% of Month 1 volume.**

---

## 8. FINANCIAL SUMMARY

| Milestone | When (Realistic) | Support Cost | % of Revenue | Action |
|-----------|-----------------|-------------|-------------|--------|
| Launch → Month 3 | Now–May 2026 | **$0** | 0% | Michael + FAQ + AI agents |
| Month 3-6 | Jun–Aug 2026 | **$50/month** (AI chatbot) | 1-4% | Automate Tier 1 |
| Month 6-9 | Sep–Nov 2026 | **$50/month** | 2-4% | Automation handles growth |
| Month 9-12 | Dec 2026–Feb 2027 | **$50-$725/month** | 2-8% | PT VA if needed for overflow |
| Month 12-18 | 2027 | **$725/month** | 5-10% | PT VA + full automation |
| Month 18+ (hiring trigger) | Mid-2027+ | **$4,883+/month** | Only if >$16K MRR | Full-time hire justified |

**Bottom line: With AI-first automation, support costs stay under $725/month for the first 18 months. Without automation, they hit $2,000+/month by Month 12. The difference is $15K-$37K over the first year — build the chatbot and FAQ early.**

---

*"The cheapest support ticket is the one that never gets created. The second cheapest is the one an AI resolves. Everything else costs real money."*