Pricing ML Consulting: From Hourly to Equity Partnerships

The Hourly Trap

I started consulting at $150/hour. It felt great for about two months. Then I realized the fundamental flaw: hourly billing punishes efficiency.

If I solve your $500K problem in 20 hours, I earn $3,000. If I take 200 hours, I earn $30,000. The incentives are exactly backwards. The client wants fast results. I want fast results. But the billing model rewards slow work.

More importantly, hourly billing caps your income at your available hours. There are only so many billable hours in a week, and some of those hours are worth 100x more than others. A 30-minute architecture decision that saves 6 months of development time is worth far more than $75.

The Evolution

Here's how our pricing evolved over 14 months:

Stage 1: Hourly ($150/hr)

• Months 1-3
• Easy to sell ("just try me for 10 hours")
• Hard to scope, hard to predict revenue
• Clients questioned every hour on the invoice
• Revenue: ~$6K/month

Stage 2: Project-Based ($5-20K per sprint)

• Months 4-8
• Fixed scope, fixed price, fixed timeline
• Clear deliverables eliminate hourly debates
• Money-back guarantee on trial sprints builds trust
• Revenue: ~$8K/month

Stage 3: Retainers ($8-25K/month)

• Months 8-14
• Ongoing relationship, continuous delivery
• Predictable revenue for me, predictable costs for the client
• Weekly demos keep alignment
• Revenue: $10K/month and growing

Stage 4: Equity Partnerships (emerging)

• Starting month 14
• ML capacity in exchange for equity or revenue share
• Long-term alignment with client success
• Only for select opportunities where ML is core to the business

How to Price a Sprint

The sprint is our bread and butter. Here's the pricing framework:

Step 1: Estimate the Value

What is the client's problem costing them? Unplanned downtime, manual labor, missed detections, false positives — put a number on it.

If a manufacturing client's quality inspection process costs $200K/year in labor and missed defects, and our system can reduce that by 60%, the value is $120K/year.

Step 2: Price at 10-20% of Annual Value

A $120K/year value problem should cost $12-24K to solve. This gives the client a clear ROI (5-10x in year one) and prices the work at its value, not its time cost.

Step 3: Scope the Sprint

4-8 weeks, specific deliverables:

• Week 1-2: Data pipeline + feature engineering
• Week 3-4: Model development + evaluation
• Week 5-6: Production deployment + monitoring
• Deliverable: Working production system with documentation

Step 4: Add the Guarantee

"Full money-back guarantee if you're not satisfied with the deliverables." This eliminates buyer risk and demonstrates confidence. In 14 months, we've never had a client invoke the guarantee.

The Retainer Pitch

Retainers follow naturally from successful sprints. The conversation goes:

"The sprint delivered [specific results]. Your system now needs ongoing monitoring, optimization, and feature development. A retainer gives you continuous ML engineering capacity at a fraction of the cost of hiring."

Retainer Pricing Math

Compare the retainer cost to the in-house alternative:

• Junior ML engineer: $120K/year = $10K/month (before benefits, equipment, management overhead)
• Our retainer: $8-25K/month for senior-level ML engineering

At $15K/month, the client gets a senior ML engineer without recruiting costs, benefits, equipment, management overhead, or the risk of the hire not working out.

What's Included

A typical $15K/month retainer:

• 80 hours of ML engineering per month
• Weekly 30-minute demo/alignment calls
• Direct Slack/email access to the engineering team
• Priority response (< 4 hour during business hours)
• Monthly performance reports
• Continuous monitoring and optimization

The Equity Conversation

Equity partnerships are the highest-leverage arrangement for both sides, but only when the fit is right.

When Equity Works

• The company is early-stage (seed to Series A)
• ML is core to the product, not supporting
• They can't afford cash retainers but have significant upside potential
• The founder is technical and values engineering quality
• There's a clear path to ML being a major value driver

How We Structure It

• 1-3% equity for 6-12 months of ML engineering capacity
• Vesting schedule aligned with milestones (not just time)
• Board observer rights for >2% equity
• Cash component for out-of-pocket costs (cloud, tools)
• Quarterly business reviews

The Key Question

Before accepting equity, I ask: "If this company raised a Series A at a $20M valuation, would my 2% ($400K) fairly compensate 12 months of senior ML engineering?"

If the answer is yes, and I believe in the team and product, the equity deal makes sense.

Lessons Learned

1. Never Compete on Price

If a prospect is choosing between you and a cheaper alternative, you've already lost — either the engagement or the margin. Compete on expertise, speed, and production quality.

2. The Guarantee Closes Deals

The money-back guarantee has been our single most effective closing tool. It shifts all risk from the buyer to us, which is exactly where it should be if we're confident in our delivery.

3. Scope Ruthlessly

The number one threat to sprint profitability is scope creep. Define what's in scope, what's out of scope, and what the change request process looks like. Write it down.

4. Weekly Demos Build Trust

Nothing builds client trust faster than showing working software every week. It's also the best insurance against misalignment — if you're heading in the wrong direction, a weekly demo catches it before you've wasted a month.

5. Retainers > Projects

Predictable revenue is worth a discount. A $15K/month retainer ($180K/year) is more valuable than three $20K projects ($60K/year) because the retainer provides consistent cash flow and the ability to plan capacity.

The Path to $100K MRR

Our pricing evolution directly maps to our revenue growth:

• Hourly: $6K/month (feast or famine)
• Sprints: $8K/month (lumpy but growing)
• Retainers: $10K/month (predictable, scalable)
• Target: $100K/month (8-10 retainer clients + equity appreciation)

Each pricing model built on the previous one. Hourly work proved our expertise. Sprints proved our delivery capability. Retainers proved our long-term value. Equity deals are the next lever.

The goal isn't to find the "right" pricing model — it's to evolve your pricing as your business matures and your track record grows.