Solution Manager · Oaxaca, Mexico · remote worldwide
Laurens Borrey.
Solution Manager.
Bring me a problem in your company. I take it apart, find where it actually goes wrong, and build the fix — then name the date and the number it gets judged on. Five years running my own companies across three countries, with my own money on the line, is where I learned to do that.
How I work
01
Measure before believing
Most urgent problems are the wrong problem. I look for the one number that settles the argument, and I go and get it before anyone builds anything.
02
Separate the symptom from the cause
What looks broken is often only measured wrong. What looks fine is often quietly losing money every day. The two get confused constantly.
03
Build the smallest fix that can be judged
A change with a bounded downside that a week can settle, so we learn something either way. Big programmes get staged the same way, so a year-long rebuild is still steerable in month two.
04
Name the date and the number
Every fix gets a review date and the metric that decides it. If it did not move, it gets reversed. I do not defend my own ideas.
Three problems, and what the analysis found
From the past year, in pricing, outbound sales and paid acquisition. The industry changes; the method does not.
“Our discounting isn’t buying us anything.”
Found
I put 28 days of automated discounting next to the 17 days before it. Sales per day had barely moved — 11.7 against 12.1 — while the average price had fallen by more than a third. The discounting was not bringing new customers in. It was selling the same inventory cheaper.
Did
Capped every discount at 15%, moved the target from volume to revenue per unit available, and set a seven-day review with the number that decides it written down in advance.
“Our outbound channel switched itself off and nobody knows why.”
Found
A delivery guard had stopped all sending for nine days over a bounce rate above its limit. The bounces were not dead customer mailboxes: our own tool had been taking whatever address a page carried, so a share of every batch went to the target company’s suppliers and agencies.
Did
Removed 161 addresses, made the tool refuse anything that is not the company’s own domain, and ranked the queue by how each address was obtained: 0% bounce from the best source against 25% from the worst. Restarted at half volume with the measurement window reset.
“We spend on ads and can’t attribute a single sale.”
Found
Thousands of clicks, not one traceable order. The campaign sent every click to a page with no checkout on it, so the tracking id was destroyed one step before checkout — not inside it, where everyone had been looking. The job feeding revenue back to the ad platform had also been failing silently on an unrelated expired credential.
Did
Built a check that matches ad click ids to real sales and sets them against what each platform claims, so budget decisions are made on traced revenue instead of reported conversions.
Where I’ve done it
I run these companies largely through AI agents I direct, which is why one person can hold this much surface area. The agents do the work. Deciding what the problem really is, and whether the fix earned its keep, is the job.
The short version
3
cities in our first 3 years — Co404 Coliving, MX & CO
$800K
raised from investors, two rounds
50+
staff across two countries
300+
Google reviews, 4.7–4.8 rating
3
legal entities: MX · CO · BE