Ultimate Beauty

Growth Had Hit a Ceiling. We Raised It 100x.

~100to10,000+

products checked per run

~7 hrsto<30 min

for the purchase plan

1to0 people

needed each morning

Talk to the Operators
Sector
eCommerce
Engagement
Operational Automation
Role
Interim CTO
Duration
2 years engagement

The Situation

Ultimate Beauty's growth was capped by how many products its team could check by hand. We automated research, purchasing and shipping, and now it checks 100x more.

A small team at their desks in a bright loft office, each working on a computer
growth the manual process could no longer support
100% a year
departments passing the same data around by copy and paste
4
of past sales, so buying decisions ran on instinct
0 records

Ultimate Beauty buys beauty products in Europe and sells them on Amazon and Walmart in the US and Canada. The business had doubled every year, and then it ran out of room. The more products it can research, the more opportunities it finds, and research had reached its limit.

Most of the company ran by hand: research, purchasing and shipping. Each morning's buying depended on one script that an engineer started before dawn. It took hours, and a failure meant starting again. The company committed capital every day on buying rules it had no way to check against past results.

Founder Matteo Lombardi wanted experienced operators inside the company, and asked us to lead its technology.

“They actually helped us build something far beyond what we thought possible.”
Matteo Lombardi Founder, Ultimate Beauty

What We Did

  1. Made the existing research run itself.

    We changed none of its logic. We set it up to start, finish and recover alone. The business had a result before anything was rebuilt.

    2 weeks to the first run that nobody had to start

  2. Rebuilt research to find new products.

    Until then, research only covered products already in the catalog. The new tool searched the market for products the company had never sold, with an AI model matching each one across Google Shopping, retailer sites and the catalog. It also kept every result, giving the company its first record of past results.

    4 data types prices, stock, demand and product details

  3. Put the company's data in one place.

    Every product's stock, price and status now sits where every tool can read it. A move to a new inventory system had stalled and was blocking that data, so we took it over. Dashboards then gave departments their first KPIs.

    1 file replaced by a unified product and data system

  4. Automated purchasing and shipping.

    Two sets of rules now make the daily calls. The first recommends which products to buy, and once a buyer approves them, they go straight into the cart. The second recommends what to ship to each marketplace, balancing stock against AI forecasts of how fast each product will sell. A person reviews both before anything is bought or shipped.

    14-day sales forecast behind every shipping plan

  5. Turned guesswork into a system that keeps getting better.

    Buying and shipping rules used to change on instinct, with no way to know if a change helped. Now every idea is tested against real past results first, using market data and the company's own sales. Only changes that prove themselves go live, and the system keeps improving.

    100% of rule changes tested on past results before going live

Before vs After

Products researched per run ~100 10,000+ 100x more
Purchase plan, time to produce ~7 hours Under 30 min 14x faster
How research starts By hand, at 5 a.m. By itself, on schedule Automated
Inventory system One failing spreadsheet One system for all data Replaced
Business KPIs No shared figures Department dashboards Full visibility
Buying and shipping rules Changed on instinct Every change tested first Optimized

Where They Are Today

2,000+ items

shipped to marketplaces every week

Research no longer stops at the best sellers. Ultimate Beauty now goes after the long tail of products, buys smarter quantities, and ships what will be selling when it arrives. Growth no longer depends on how much the team can examine by hand. Its own team still runs and improves the system.

When we arrived 100% a year growth the manual process could no longer support

What SIROC Brought

We led technology as Interim CTO, and we started with the business. We learned how research, purchasing, operations and shipping worked before changing any of them. Then we built them into one supply chain, each step feeding the next, with sales forecasts behind both the buying and shipping recommendations.

A full rebuild would have meant a long wait for the first result. We improved what existed step by step, so the company kept buying every day while its systems changed underneath.

“By the end it was a completely new company: fully automated, and able to keep building on the data and automation system we created.”
Javier García Interim CTO, SIROC
leading technology
Interim CTO
taken over, though it was outside our brief
Migration
that predicts how fast each product will sell
Forecasting
the company kept buying while its systems changed
Every day
The Ultimate Beauty team waving under canvas shades in the desert at dusk
A buying rule tested on 12 months of past sales, passed and ready to go live
A research run that checked 10,482 products, listing new opportunities with their AI match and ROI

When growth depends on how much a team can check by hand, growth stops. AI-powered research and forecasting took that limit away.

  • ~100 → 10,000+ products checked per run
  • ~7 hrs → <30 min for the purchase plan
  • 1 → 0 people needed each morning