LAB · ACTIVE
EST 2026 · V 1.1
001 / HOME
A PERFORMANCE-MARKETING LAB

Limits are
untested
assumptions.

We write the hypothesis, ship the test, and let the result decide what happens next.

Experiment · Optimize · Go beyond
// HOW WE WORK
01METHOD
Hypotheses before spend.
Every test gets a prediction, an MDE, and a sample size before a dollar runs. Forming the hypothesis after you've seen the data isn't strategy — it's storytelling.
02DISCIPLINE
Kill fast.
A losing variant doesn't get a second look. Nursing a dead test is the same as not running the next one. Most tests lose; the math only works if you act on it.
03MEASUREMENT
Prove it, don't claim it.
Most marketing dashboards reward the channel that showed up last. We measure what only happened because of the change — not what got credited to it.
002 / ORIGIN
ON THE NAME

Non Plus Ultra.
Someone struck the Non.

The Pillars of Hercules marked the edge of the known world. The inscription warned sailors: nothing further beyond. This is the end.

Charles V struck the Non. What remained was Plus Ultra — further beyond — and the motto of an empire that planned to act like the warning didn't apply.

We took the name because it describes the work. We go beyond.

// LAB AXIOM 001
"The map ends where the last person stopped looking."
— posted above the desk
003 / METHOD

Three steps. One ascent.

Every cycle runs the same loop: find the move, compound it, push past the ceiling it revealed.
Experiment · Optimize · Go beyond
// THE LIFT CURVE · METHODOLOGY VS CONTROL
CONTROL 01 02 03 · GO BEYOND LIFT EXPERIMENT OPTIMIZE GO BEYOND TIME →
// REPRESENTATIVE METHODOLOGY · ENGAGEMENTS VARY
// THE LOOP
01EXPERIMENT

Design the test.

A hypothesis on paper before any traffic moves. Tests are small enough to run in a week, rigorous enough to read in two.

  • Pre-registered hypothesis
  • Power-calc'd sample size
  • Control + treatment isolated
  • Cost ceiling defined upfront
02OPTIMIZE

Refine the winners.

Most lifts come from the second iteration, not the first. Once we've isolated what moves, we compound it — variant by variant — until the curve flattens.

  • Daily reads, weekly decisions
  • Creative refresh on signal, not schedule
  • Channel mix re-balanced on signal
  • Nothing scaled without proof
03GO BEYOND

Ship what beats control.

Every cycle ends with analysis of what we ran, what beat control, what didn't, and where the curve says to push next.

  • Readout: wins, losses, the middle
  • Winners become the new control
  • Losses logged for the next test
  • Last cycle's ceiling becomes next cycle's floor
// LAB AXIOM 002
The cost of running
the test is always
less than the cost of
not knowing.
— THE LAB · OPERATING PRINCIPLE Plus Ultra Labs · A performance-marketing lab
004 / ABOUT

A lab,
not a guru.

A lab that only reports its wins isn't a lab. It's marketing for itself. We report all of it — winners, losers, and the boring middle. Plus ultra is only a discipline if the readout is honest. You can't push past a ceiling you haven't measured.
004.A
ORIGIN

The name is a dare, not a flex.

Every limit in this work is somebody's Non Plus Ultra — a warning carved into the map by someone before you.

Every "this won't scale," every "we already tried that," every "channel ceiling" is the same inscription in different paint. The job is to design the test that would settle it and let the result decide whether the warning held.

Not reckless energy — the opposite. The discipline of someone who'd rather be embarrassed by a clean negative than comfortable with an unchecked assumption.

// LAB AXIOM 003
Every strategy
is a hypothesis.
Every delta is a verdict.
We only ship the second.
— THE LAB · OPERATING PRINCIPLE Plus Ultra Labs · A performance-marketing lab
004.C
DISPOSITION

How we think — and how we don't.

If you're still reading after the right column, we'll get along.

+ How we think

  • The control is the enemy. Beating it is the job.
  • Hypotheses get written before the test, not after the win.
  • Statistical significance is a floor, not a finish line.
  • Losing tests are data — we don't bury them.
  • Incrementality beats attribution.
  • The readout is the deliverable. Everything before it is setup.

− How we don't

  • "Directionally positive."
  • Peeking at running tests and calling it pacing.
  • Attribution models built to flatter a channel.
  • Optimism as a methodology.
  • Frameworks named after agencies.
  • Reporting that needs a translator.
  • Confidence as a deliverable.
004.D
FOUNDER

About the founder.

Plus Ultra Labs mark

Founder

Plus Ultra Labs

More than 20 years across software, product, and growth. Former senior developer who moved into product, leading CRO — because the only honest way to know what works is to test it against a control. 2,300+ experiments run. Multi-million in revenue lift across digital products. Technical depth where most marketers wave hands; product instinct where most engineers shrug.

005 /
CONTACT
// Direct
hello@plusultralabs.co
// Mail
1388 Haight St, Ste 467
San Francisco, CA 94117
Referral inquiries welcome Response time varies