Est. reading time: 5 minutes

This case study covers the email automation layer of the Beauty by Amanda Earnest engagement. The full growth story, including the Meta and TikTok system that fed it, is here.

The Situation

Amanda Earnest sells color-matched foundation as a solo Seint Beauty artist, and the product is the match itself. Every lead submits a Typeform color match request, and every request gets a personal read on skin tone, undertone, and finish preference, delivered by email as a specific product recommendation.

When the engagement began in early 2020, Amanda wrote each of those emails by hand. A single match took about ten minutes, and her capacity topped out around 20 to 30 leads a day before quality started to slip. Past that point, every new lead was a lead waiting in a queue.

The plan was to grow the business through paid social. At hand-written speed, that plan had a hard ceiling on day one, because acquisition can only run as big as the response system behind it.

The Primary Challenge

The obvious fix was templates, and templates would have gutted the business. A color match is worth paying for because someone with a trained eye actually looked. A templated match reads templated, and the moment a recipient senses a form letter, the recommendation loses the authority that closes the sale.

Headcount was off the table too. Amanda ran the operation alone inside a direct sales ecosystem, and the economics of the business were built around one operator.

So delivery had to get faster by an order of magnitude without the output ever reading as automated. Speed at the cost of the personal read would have cleared the backlog by destroying the reason leads showed up at all.

Key Outcomes

• Per-match handling time cut from roughly ten minutes of writing to thirty seconds of selection
• Daily capacity expanded from 20 to 30 leads to hundreds, with a personalized recommendation on every match
• 42,000+ color match requests handled by one operator across five years
• Automation rebuilt repeatedly as products, pricing, and matching logic changed, without interrupting live lead flow
• System supported a business generating $100,000+ in monthly revenue at peak

Our Approach

We drew the line at judgment. Every match stayed hers. What we automated was everything after the decision. The writing, the product references, the assembly of the email she would have typed by hand.

Her recommendation patterns became selectable attributes, and the email became a system that assembled itself around her selections. The recipient still got Amanda’s read. Delivering it no longer cost her ten minutes.

Execution Highlights

A Drop-Down Menu That Wrote the Email

The core build was a Mailchimp automation driven by attribute selection. Amanda reviewed a color match request, chose the matching attributes from a drop-down menu, and the system auto-populated a fully personalized recommendation email around her selections.

The drop-down selections fed Mailchimp merge fields and conditional content blocks, turning Amanda’s product choices into a finished recommendation without requiring her to rewrite the same explanations on every match.

Ten minutes of writing became thirty seconds of selection, and the email on the other end stayed as specific as the hand-written version. Same products, same reasoning, same voice. The only thing removed was the typing.

Structured Intake Through Typeform

Every lead arrived through a Typeform color match request, so the details Amanda matched against were already organized before she opened the submission. Skin tone, undertone, and finish preference came in the same structure every time, which let her make the match fast and gave the email system the same attributes to assemble around. That shared structure is what made thirty-second turnarounds hold at volume.

Rebuilt in Motion, Every Time the Catalog Moved

The system never reached a finished state. Seint’s products changed. Pricing changed. Amanda’s matching logic evolved as the catalog grew. Each shift forced a rebuild of the automation to keep the recommendations current, and every rebuild happened under live lead flow, because the business could not pause while its infrastructure caught up.

Over five years, the setup was expanded again and again to stay ahead of growth that never waited for it.

The Standard Was Invisibility

The automation had one pass-fail test. If a recipient could tell the email was automated, it failed. Every rebuild was held to that standard, which is why the system never drifted toward template even as it scaled past anything hand-writing could have handled.

Results

The operational numbers are the results. Ten minutes became thirty seconds. Twenty to thirty matches a day became hundreds. One operator handled 42,000+ color match requests across five years, each answered with a personal recommendation, and at peak the system supported a business generating over $100,000 in monthly revenue.

The strategic result is what those numbers unlocked. With the throughput ceiling gone, paid social could finally run at full volume. Meta opened the funnel in 2020, TikTok multiplied it in 2022, and every lead they produced landed on a back end built to absorb it. That acquisition story, including the cost per lead falling from roughly $3 to 90 cents, is covered in the full engagement case study.

Constraints We Navigated

None of this happened in clean conditions.

• A solo operator structure, so every workflow had to run without added headcount
• Seint’s direct sales structure and rules, which put some standard ecommerce mechanics off the table
• Live lead flow throughout, so every rebuild happened with the engine running

Strategic Takeaway

When the product is a personalized response, the growth constraint is response capacity, and most operators try to break that constraint with the wrong tools. Templates flatten the value. Headcount scales the cost. The durable fix is a system that assembles the operator’s expertise fast enough to keep pace with acquisition.

That is what this build did for five years. Amanda made every call, the system did every keystroke, and 42,000+ requests got a recommendation that read like she wrote it herself. Automate the delivery, never the judgment.