Pricing Psychology for Solo Founders — Let AI Help You Find the Optimal Price

In September 2025, JewelFlow went from $49/month to $69/month — a 40% increase.

I was nervous for two solid weeks before the price change. JewelFlow had 180 paying customers at the time. I did the math: if 20% of customers churned due to the increase, monthly revenue would actually decline. This was not a decision I could afford to wing.

The results after the increase: 6 customers churned within 30 days (3.3%), new customer acquisition showed no noticeable change, and monthly revenue went from roughly $8,820 to approximately $12,006. Three months later, the customer base stabilized at 190+.

This wasn't luck. Before the price increase, I ran 6 rounds of pricing experiments over about two months, systematically validating "how much customers are willing to pay for JewelFlow."

This article breaks down the entire pricing methodology: how to analyze competitor pricing, how to test willingness to pay, how to design A/B tests, and what role AI plays at each step.

---

Background: Common Pricing Mistakes Solo Founders Make

Most independent founders price their products by looking at what competitors charge and picking a similar number.

The problem with this approach: you don't know if the competitor's pricing is right either. Many competitors set their prices by gut feel too. Following a potentially wrong anchor means two people getting lost together.

Another common mistake is pricing too low. I've seen far too many cases of this in the Solo Unicorn Club — a member builds a great product, prices it at $19/month because they're "afraid nobody will buy it if it's more expensive." After doing proper research, they discover customers are perfectly willing to pay $49, even $79. Every $30/month you're leaving on the table, multiplied by 100 customers, is $36,000 less per year.

JewelFlow's original $49/month price came from the same psychology — I assumed small and mid-size jewelers had tight budgets. The data later showed me that JewelFlow saves them 5–8x the monthly fee in labor costs. At $69, it still registers as "cheap" in their value perception.

---

Core Approach: Three Principles

Principle 1: Price Based on Value, Not Cost

Cost-based pricing logic goes: my server, API, and labor costs add up to X, I add a profit margin, and I price at Y.

The problem is that customers don't care about your costs. They care about "how much money you save me" or "how much more money you help me make."

JewelFlow's server and API cost per customer is under $5/month. Under a cost-plus model, $15–$20 might seem reasonable. But JewelFlow saves a mid-size jeweler roughly $300–$500 per month in labor — automated quoting, customer follow-ups, inventory management. At $69/month relative to that value, customers feel they're getting a deal.

How do you quantify value? I used Claude for a simple but effective analysis:

1. Collected usage data from 30 active customers — how many quotes they processed through JewelFlow, how many manual steps they saved
2. Fed that data along with average jewelry industry labor costs ($18–25/hour) to Claude for calculation
3. Output: estimated monthly value saved per customer through JewelFlow

The median was $380/month. This means the $69/month price captures only 18% of the value customers receive. That ratio is actually on the low side for SaaS — 25%–30% is considered healthy.

Principle 2: Validate Intuition with Data — Don't Substitute Data with Intuition

I didn't just calculate value and make the call. The numbers told me "you can raise prices," but how much and how — that needed testing.

I conducted three types of data collection:

Competitive pricing analysis: Used Claude to scrape and analyze the pricing pages of 8 JewelFlow competitors — Stuller's enterprise tools, JOOR, Valigara, and others. Claude produced a structured comparison: feature coverage vs. price, per-user tiers vs. feature tiers, annual billing discounts. Conclusion: the mid-range for comparable tools was $59–$99/month, and JewelFlow's $49 was clearly on the low end.

Willingness-to-pay survey (Van Westendorp): This is a classic pricing research method. I used Tally to create a simple survey and sent it to 50 existing customers with four questions:

1. At what price would you consider the product "too cheap" and question its quality?
2. At what price would you consider the product "not expensive — acceptable"?
3. At what price would you consider the product "starting to get expensive, but you'd still use it"?
4. At what price would you consider the product "too expensive and would stop using it"?

32 customers responded. Claude ran the cross-analysis, producing an "optimal price range" of $59–$79/month, with the median "acceptable price ceiling" at $89/month.

Small-scale A/B test: New sign-ups were randomly split into two groups — Group A saw a $59/month pricing page, Group B saw $69/month. The test ran for 6 weeks with 45 users in each group. Conversion rates: Group A 14.2%, Group B 13.1%. The difference was not statistically significant, meaning $69 vs. $59 had minimal impact on conversion.

All three data sources pointed to the same conclusion: raising prices to $69/month was safe.

Principle 3: Be Transparent About Price Increases — Give Existing Customers a Soft Landing

Setting the price isn't enough — the "how" of a price increase matters just as much as the "how much."

My strategy:

• Existing customers locked in at the old price for 3 months, with the new price taking effect in month 4
• The price increase email clearly explained why — listing 4 core features added in the past 6 months
• Offered an annual plan: $59/month ($708/year), giving customers a "15% below the new price" off-ramp
• For the highest-usage customers (top 10% by monthly usage), I reached out individually with custom arrangements

Results: Of the 6 churned customers, 4 were low-frequency users (likely on their way out anyway), and 2 explicitly said "the budget is genuinely tight." Zero customers left negative reviews or feedback due to the price increase.

---

Tool Stack Breakdown

Use Case	Tool	Monthly Cost	Why I Chose It
Competitor pricing page monitoring + analysis	Visualping + Claude API	~$12	Monitors competitor price changes + structured comparison
Willingness-to-pay survey	Tally (free tier)	$0	Clean, intuitive, free
Survey data analysis	Claude API	~$3 (one-time)	Van Westendorp cross-analysis
Pricing page A/B testing	PostHog (free tier)	$0	Open source, self-hosted, free up to 1M events/month
A/B test implementation	Stripe + custom script	$0 (no extra cost)	Already using Stripe; used metadata to differentiate test groups
Customer value quantification	Claude API + JewelFlow data	~$2 (one-time)	Calculated per-customer savings value
Price increase email delivery	Resend	$0 (within free quota)	100 emails/day free
Total		~$15–17/month

Pricing experiments aren't an ongoing cost — most tools are used once or on-demand. The total tool spend during the two-month intensive testing period was under $50.

Why not a dedicated SaaS pricing tool (PriceWell, Intelligems)? PriceWell starts at $49/month; Intelligems targets e-commerce. At my scale (under 200 customers), the PostHog + Stripe + Claude combination is more than sufficient, and the money saved is worth more than the extra features those tools provide.

---

Real Numbers

Key metrics before and after the JewelFlow price increase:

Revenue:

• Pre-increase monthly revenue: ~$8,820 (180 customers x $49)
• 1 month post-increase: ~$12,006 (174 customers x $69)
• 3 months post-increase: ~$13,110 (190 customers x $69)
• Annualized growth: ~$51,480 to ~$157,320 (including contribution from continued new customer growth)

Customer Churn:

• 30-day churn rate post-increase: 3.3% (6/180)
• Normal monthly churn rate: 2.1%
• Net incremental churn attributable to pricing: ~2 customers

Customer Perception:

• NPS post-increase: 56 (pre-increase: 58, essentially flat)
• Price increase feedback emails received: 11 — 7 positive ("Understood, the product really has been improving"), 4 neutral ("Noted"), 0 negative

ROI of the Pricing Experiments:

• Total experiment investment: ~$50 in tool costs + ~30 hours of personal time
• Monthly revenue increase from pricing change: ~$3,186
• Payback period: a day or two

---

Lessons from Mistakes

Mistake 1: Drawing conclusions from insufficient A/B test sample size.

My first A/B test only ran for two weeks with 15 users per group. The conversion rate difference looked dramatic (Group A 20%, Group B 6.7%), but it was statistically meaningless — the sample was too small, and random noise alone could produce that gap. I nearly abandoned the price increase because of it.

After extending to 6 weeks and 45 users per group, the difference narrowed to 14.2% vs. 13.1%, well within normal range. Lesson: pricing A/B tests need at least 40–50 valid samples per group, or the data is useless.

Mistake 2: Survey wording skewed results.

The first version of my Van Westendorp survey asked "How much do you think this product is worth?" — a phrasing that nudges respondents toward lower numbers. After switching to the standard four-quadrant questions, the results were completely different, with the optimal price range shifting up by $15. Wording details make an enormous difference in pricing research. I recommend sticking to the original phrasing from established methodologies.

Mistake 3: Only comparing competitor prices while ignoring their value narratives.

Initially, my competitive analysis only compared price numbers. After Claude ran a deeper analysis, it pointed out that Competitor X's pricing page emphasized "saving time" while Competitor Y emphasized "increasing sales" — same type of product, but different value narratives led to customers being willing to pay more than 2x for one over the other.

This insight directly influenced JewelFlow's post-increase pricing page redesign. I changed the core messaging from "Automate your jewelry business management" to "Serve 3x more customers every month" — emphasizing revenue growth rather than efficiency gains.

Mistake 4: Springing the price increase on existing customers.

I originally planned to send one email and start the new pricing immediately. A member of the Solo Unicorn Club with SaaS experience reminded me: give at least 30 days' notice and a buffer period. I took the advice and added a 3-month lock-in period. In retrospect, this was one of the key factors in keeping churn at 3.3% — customers had time to process and adapt, rather than being hit with a sudden change.

---

Advice for Getting Started

Step 1: Calculate how much value your product creates for customers.

If you don't know how much value customers get from your product, you can't price it rationally. Spend an afternoon using Claude to analyze your customer data and calculate "how much money your product saves or earns for each customer per month." That number is the ceiling for your pricing.

Step 2: Run a Van Westendorp survey.

Four questions, built in Tally in 5 minutes. Send it to 30–50 existing customers. No fancy statistical tools needed — just feed the results to Claude for analysis. You'll get a clear "optimal price range."

Step 3: Test small, don't go all-in.

A/B testing on new users is much safer than raising prices on existing customers. First validate price acceptance with new users, then consider your strategy for existing customer increases once you have confirmation.

---

Final Thoughts

Pricing is the most underrated growth lever for solo businesses. Customer acquisition costs keep rising, conversion rates plateau, features max out — but a 30% price increase might contribute more to revenue than all of those efforts combined.

But you can't price by feel. AI now enables a single person to do systematic pricing analysis — competitive benchmarking, willingness-to-pay research, A/B testing — work that used to require pricing consultants ($5,000–$20,000 per engagement) can now be done for under $50 in tools.

$69 vs. $49 — it looks like a $20 difference. Multiply by 190 customers, multiply by 12 months, and it's $45,600/year. That's the gap between scientific pricing and guesswork.

Is your current price calculated, or guessed?