Derek Hales on Building NapLab's Mattress Test Operation

Derek Hales on Building NapLab from an Apartment to a Testing Operation

NapLab began with a shopping problem and grew into a physical testing operation. In episode 100 of Venture Step, founder Derek Hales explains what happened between those two points: too many king mattresses in an apartment, years of publishing, dedicated space, repeatable tests, scoring rules, and the continuing work of showing readers how a recommendation was produced.

The durable lesson is not that every review needs expensive laboratory equipment. It is that a review becomes more useful when another person can inspect the product, method, evidence, incentives, and limits behind the conclusion.

The apartment was the first operating constraint

Hales traces his entry into mattress reviewing to 2014. He and his wife, Samantha, were shopping for a mattress and found the store options expensive. They tried mattresses sold online, documented the experience, and published what they learned.

The work quickly became physical. Hales recalls keeping as many as 17 king mattresses in a one-bedroom apartment. After moving into a house, the inventory reached roughly 47. Those numbers come from his interview account, not an independent inventory record, but they reveal a problem that disappears from a finished review page. A physical product reviewer needs a place to receive, open, test, compare, retain, and eventually remove products.

Space changes the editorial possibilities. A product that is returned immediately cannot be retested after a scoring change. A crowded room makes consistent camera position, temperature, setup, and movement harder. Storage policy can determine whether a reviewer can compare an older product against a replacement rather than relying on memory.

Hales says usable mattresses that are no longer needed are donated to a local charity, while important models may be retained for future comparison. That decision belongs to the operating system behind the content. It affects cost, space, repeat testing, and the evidence available when a reader asks whether a new model is actually different.

NapLab's current about page says the operation now uses a dedicated Scottsdale testing studio and storage space. It also says the team grew from one person in an apartment to nine people. Those are current first-party descriptions as of July 27, 2026, not independently audited staffing or facility records.

flowchart LR
    A["Shopping problem"] --> B["First comparison published"]
    B --> C["More products and reader questions"]
    C --> D["Repeatable tests and records"]
    D --> E["Dedicated testing and storage"]
    E --> F["Versioned scores and corrections"]
    F --> G["A recommendation readers can inspect"]

The diagram makes the real progression visible. Publishing was the beginning. The method, facility, records, and correction process made comparisons more durable.

A simple test can still carry evidence

Several of Hales's early methods were deliberately easy to understand. Sinkage could be observed by placing a medicine ball on a mattress, photographing it from a controlled position, and using the image to estimate depth. Motion could be tested by dropping a ball on one side and measuring what happened elsewhere.

The important distinction is not cheap equipment versus expensive equipment. It is whether the procedure defines the object, position, force, instrument, environment, observation, and calculation well enough for a repeat to mean something.

NapLab's current testing and scoring methodology is more developed than the recording-era examples. The page, updated May 18, 2026, identifies its current scoring system as version 1.3. It describes eight weighted scoring factors, as well as non-scoring measurements such as sinkage, bounce, foam density, and coil gauge. Durability now has its own score but is not yet included in the overall score while the company gathers more data.

That separation is editorially useful. NapLab says it does not score sinkage because more or less sinkage is a preference, not a universal improvement. It also describes pressure relief as a structured subjective assessment informed by experience, construction, pressure mapping, and other observations. A number can be measured consistently without proving that a larger or smaller number is better for everyone.

The methodology also contains an errors and corrections section. NapLab acknowledges that tests, calculations, and updates can be wrong and invites readers to report problems. A correction policy does not establish that every result is accurate. It does make error handling part of the public method rather than an invisible exception.

Repeatability is narrower than truth

Hales uses the word objective to describe much of NapLab's work. That word needs a boundary.

A measurement can be repeatable under the same conditions and still omit a relevant part of the product experience. It can also change when the operator, day, environment, sample, or instrument changes. The NIST gauge study guidance separates repeatability, reproducibility, stability, bias, resolution, linearity, hysteresis, and drift. Its measurement variability guidance warns that day-to-day variation may exceed the short-term precision of an instrument.

This does not mean a small review operation must behave like a national metrology institute. It means a credible method should say what was held constant, what may have varied, and what the observation can support.

NapLab's durability test illustrates the boundary. Its current page says a rollator moves across a mattress 100,000 times over about five days and is intended to simulate ten years of use. The company says it aims to follow ASTM F1566 but does not follow every requirement. NapLab is also explicit that ASTM does not endorse or certify its work.

The appropriate public statement is therefore precise: NapLab runs an accelerated rollator procedure and interprets changes in height, sinkage, and firmness as durability evidence under that method. It is not a promise that one buyer will receive exactly ten years of service. Real use introduces bodies, foundations, rooms, handling, manufacturing variation, and other conditions that the test does not reproduce.

A score is a model, not a property inside the mattress

The episode turns repeatedly to the attraction of a single number. A total score makes comparison faster, but it also compresses measurement, judgment, normalization, weights, and version history.

NapLab's version 1.3 page explains that its overall score is a weighted average of eight factors. The company score within that model is currently built from returns, trial period, warranty, and shipping, with its own internal weights. The site also says its best-mattress selections consider other factors beyond the test score, including materials, value, availability, specialization, brand trust, and subjective assessment. The current selection-method page says final recommendations remain a team decision and are reviewed periodically.

That is not a flaw hidden behind the score. It is the nature of a recommendation. Test data can establish comparative observations. A ranking still has to decide which observations matter, how much they matter, and to whom.

The OECD and Joint Research Centre handbook on composite indicators was written for policy indicators, not mattress reviews. Its methodological warning transfers carefully: indicator selection, normalization, weighting, aggregation, and missing values can change a composite result. Sensitivity analysis helps reveal whether a ranking survives reasonable alternate choices.

For a buyer, the practical question is not simply which mattress has the highest total. It is whether the measurements and priorities inside the score resemble the buyer's own decision.

Affiliate revenue makes governance visible

NapLab earns money through referral links. Its about page says a purchase through those links may generate a commission. More unusually, the company says those relationships may influence which products it reviews and writes about, as well as where products appear. It says referral relationships do not affect scores and that brands cannot pay for a review or a higher rank. The same page publishes a product-level table showing referral-link availability and whether a mattress was supplied free.

Those statements are evidence about NapLab's declared policies. They are not an independent audit of how every editorial decision was made.

The distinction matters because disclosure and method solve different problems. The FTC's current affiliate guidance says a material commission relationship should be disclosed clearly and conspicuously, close enough to the recommendation and link for readers to understand it. The FTC also says labels such as "affiliate link" may not communicate that the publisher earns money.

A clear disclosure helps a reader interpret incentives. It does not validate a test, prove that selection was neutral, or show that negative results receive the same visibility as positive ones. Those questions require public rules for product selection, access, scoring, updates, corrections, and commercial placement.

Proof of work is the useful trust signal

The strongest part of Hales's story is not the scale of the equipment or the number of products. It is the movement from an opinion that asks for trust to a method that exposes more of the work.

A reader should be able to identify the exact product version, see how it was obtained, understand the test conditions, distinguish measured and subjective findings, inspect the scoring version, learn how the publisher earns money, and find the limits or corrections. The result can still be disputed. That is precisely why the evidence must remain visible.

This standard applies beyond mattresses. In [[Episode Story - Josh Sprague and the Standard Beyond Good Enough|Building Better Products by Refusing to Accept Good Enough with Josh Sprague]], physical testing reveals problems that a specification sheet cannot. [[Logan Yonavjak on Measuring Founder Readiness|Measuring Leadership Capacity with Logan Yonavjak and Readiness Engine]] raises the related problem of turning observations into governed assessments. [[E099 Content Plan|Andy Ellwood and Stretch on Grocery Price Transparency]] shows why consumers need access to the inputs behind a confident commercial claim.

The next useful step is not to accept or reject NapLab as a whole. Read [[How to Build a Repeatable Product Testing Protocol]], then apply [[How to Read Product Reviews Without Getting Sold]] to one review you are considering. The method should earn influence claim by claim.

About Derek Hales

[[Derek Hales Guest Profile|Derek Hales]] is the founder and editor-in-chief of NapLab. His current first-party biography says he began mattress testing in 2014, founded Sleepopolis that year, sold it in 2017, later built Modern Castle, and launched NapLab in August 2021. Episode 100 focuses on the operating choices behind his review work rather than recommending a particular mattress.

Sources and editorial notes

This article was written from the preserved episode 100 transcript and checked against NapLab's current methodology, company and revenue disclosures, and Derek Hales biography. Independent context comes from NIST measurement guidance, the OECD composite-indicator handbook, and the FTC Endorsement Guides Q&A.

Facility, staffing, test counts, ownership, policies, and founder history are attributed to NapLab where independent confirmation was not necessary or available. No current mattress ranking, medical outcome, life-span guarantee, or ASTM certification is asserted. AI assisted with research organization and drafting under editorial review. Publication remains unauthorized.