Skip to content

The Roastery Lab · Research

Roasting, not vibes

Every question becomes a controlled experiment — hypothesis, control, one variable, a cupped conclusion.

These are open questions, not answers. Pacific Roots is pre-launch, so every experiment below is designed and queued and not one has run yet — you are reading the Lab's research agenda in the open, before the first batch drops.

The method

Nine steps from question to published finding

A finding that confirms the hypothesis and one that overturns it are equally the goal — the only failure is a question too vague to answer.

  1. 01

    Define the question

    Start where the cup does — a flavor we chase, a batch that surprised us, a claim we refuse to take on faith.

  2. 02

    Write it falsifiable

    State the hypothesis so it can be proven wrong. 'A longer rest reads sweeter' beats 'rest matters' — one we can actually test.

  3. 03

    Control + one variable

    A reference roast holds the line; a single dial moves. Change two things at once and you have learned nothing you can name.

  4. 04

    Lock the constants

    Everything else is fixed and written down first — green lot, curve, grind, water, ratio — so the result belongs to the variable, not to luck.

  5. 05

    Link batches & capture

    Every roast logs to its batch ID: curve, events, weight loss — captured the same way, every time, before anyone tastes.

  6. 06

    Collect the outcomes

    Cup blind on the seven axes; pull the brews; read TDS and extraction. Sensory and numbers, side by side.

  7. 07

    Analyze with its limits

    Read what the data says and, out loud, what it cannot — small n, one origin, a single palate. The caveats are part of the finding.

  8. 08

    Decide for production

    Confirmed, changed, or held: the result either moves a house profile or leaves it exactly where it was, on purpose.

  9. 09

    Publish it either way

    A clean answer and an inconclusive one both get written up. A question honestly settled is the win; 'not yet' is a finding too.

The open queue · nothing run yet

Four questions the cup keeps raising

Each is a falsifiable hypothesis still waiting for its control batch. We label them open because that is exactly what they are — designed, queued, and unrun.

Open hypothesis

Queued

Same final color, different development time

Which carries more sweetness into the cup?

Open hypothesis

E-001 · first to run

Same curve, different rest window

Where does each roast actually peak?

Open hypothesis

Queued

Same lot, different charge temperature

How much does the first 90 seconds decide?

Open hypothesis

Queued

Different roast degree, espresso vs filter

One coffee, two systems — where do they diverge?

E-001

The first original dataset · planned, not run

The rest-window study

Does the same curve peak at different rest days by roast level?

E-001 will be the first experiment the Lab runs end to end. One reference roast, one locked curve, cupped blind at days 6, 8, and 10 off roast — the same batch tasted across its rest window — to find where each roast level actually peaks. Every Pacific Roots bag will name a rest window; this is where that number stops being folklore and becomes ours.

Hypothesis

Same curve, different rest — each roast level peaks on a different day.

Control

One reference roast, one locked curve, one green lot.

Variable

Days off roast — cupped at 6, 8, and 10.

Capture

Blind cuppings linked to a single batch ID, seven axes each.

Reference windows

Filter 4–8 d · espresso 6–12 d · cold brew 5+ d.

Output

A published rest curve — the peak day per roast level.

Status · designed, not yet run — the finding will publish to Evidence

Where findings will publish →The rest science →

← The Roastery Lab·Confirmed or not, the answer gets published