How we know what we publish

Every infusion guide on this site started in our own kitchen, not on a content brief. Decarb curves, butter strain-loss percentages, tincture potency drift over weeks, edible onset timing — these are recorded across multiple batches with the same flower, the same scale, and the same thermometer before we ship a page.

We do not own a lab HPLC. When we cite an extraction efficiency or a cannabinoid potency number, it's either (a) drawn from published research and labeled as such, or (b) our kitchen estimate based on starting material and finished volume, labeled as such. We will never present a kitchen estimate as a lab number.

When a claim needs a citation, we prefer sources in this order:

1. Peer-reviewed primary research (PubMed, NIH, Journal of Cannabis Research, etc.) — for mechanism and pharmacokinetics claims.
2. Government and standards bodies (FDA, NIDA, EFSA, USDA, state cannabis control boards) — for safety, legality, and regulatory claims.
3. Recognized industry references (Leafly Science, Project CBD, university extension publications) — for context and accessible summaries.
4. Our kitchen testing — always labeled "in our testing" or "we found." Never dressed up as research.
5. Anecdotal community knowledge — labeled as such, used sparingly, and only when no better source exists.

If a page makes a numeric claim without a citation or a "we measured this" hedge, treat it as a bug and tell us.

BeTheHippy is reader-supported. When you buy through a link on this site we may earn a commission, at no extra cost to you. We run affiliate links with Amazon Associates, Awin, CJ, ILGM, and a small number of direct retailers.

What that does and doesn't change:

• We never accept paid placements. No brand has ever bought a recommendation on this site, and no one ever will.
• We recommend the product first, then look for an affiliate program. If the best tool for the job has no program, we still recommend it — sometimes with an external link that pays us nothing.
• If we've tested it, we say what we didn't like. Honest cons, "skip this if…" notes, and cheaper alternatives appear next to the recommendation.
• Every affiliate link is disclosed at the point of click — not buried in a footer page.

Every flagship guide carries a Last reviewed date in its header. We re-review evergreen content at least once every twelve months and immediately when (a) a cited source changes, (b) a reader flags a contradiction, or (c) we learn something in the kitchen that changes the recommendation. Material updates are logged in the public changelog.

If a page's Last reviewed date is more than a year old, treat the numeric claims with extra skepticism and email us to nudge it up the queue.

These are the lines we will not cross, regardless of what an SEO consultant or affiliate manager tells us would convert better:

• We don't make medical claims. Cannabis is not our prescription pad.
• We don't publish fabricated review counts, ratings, or "as seen in" badges.
• We don't use "studies show" without naming the study.
• We don't recommend products we haven't used or thoroughly researched.
• We don't sell flower, edibles, or any consumable. Ever.
• We don't sell, rent, or share your data with anyone outside the platforms we name in our privacy policy.

The Machine Finder quiz uses rule-based weights, not a black box. Each of the six questions adjusts each machine's score by a fixed amount — these are the weights, in full, so any recommendation we make is auditable:

The break-even calculator uses one canonical retention pair: manual 55%, machine 75%. These are community-test averages for properly-run cycles; user error widens the range. Break-even batches are derived as:

extra mg recovered per batch = flowerGrams × 1000 × THC × (0.75 − 0.55)
$ recovered per batch          = flower $ per batch × (0.75 − 0.55) / 0.55
break-even batches             = ceil( machine cost / $ recovered per batch )

The same equation is printed inline by the calculator's "Show your work" toggle with your specific numbers substituted in. If your kitchen measurements diverge materially, please tell us via the corrections form — we'll update the assumptions and log the change.

Every page tagged Hands-on tested ran through this protocol. Pages tagged Based on specs + owner reports did not — we're honest about which is which so a spec-sheet summary isn't dressed up as a first-person test.

The Manual Cannabutter Setup Test is our first canonical hands-on run:

• Fixed inputs. Same strain, same weighed flower quantity (7g), same butter brand and weight, same oven-thermometer-verified decarb, same stovetop with a calibrated candy thermometer. Substitutions get their own logged run.
• Fixed process. Mason-jar decarb → double-boiler infusion at 160–180°F for the documented time → cheesecloth strain → weighed yield.
• Logged in the batch tracker. Every field on the printable Batch Tracker PDF (dose, decarb temp/time, infusion temp/time, finished ml, assumed efficiency, mg/serving) gets filled in during the run — no reconstruction after the fact.
• What we photograph. Setup shot, decarb-in-jar, infusion-at-temp, strain step, finished butter, filled-in tracker page. No stock imagery in hands-on posts.
• What disqualifies a run. Missing tracker field, out-of-range decarb temp, unweighed flower, forgotten photos. A disqualified run is documented and re-run; it doesn't get published as evidence.
• Verdict rules. "What worked" is what a first-time home cook would repeat. "What we'd skip" is what didn't earn its spot. We don't compare to machines we haven't independently tested in the same session.

Grab the same tracker we use — it's the second PDF in the free cheat sheet + batch tracker bundle. Log every batch and your third one will be dialed in.

Found a claim that doesn't add up? A source link that's broken? A recommendation that didn't match your kitchen results? That's exactly the feedback that makes the site better. Use the corrections form or email us directly — every flagged item gets a response and, when warranted, a logged change in the public changelog.