How we pick, rate and track prices

mid1 publishes 5 buying guides and 140 product pages, written by one named reviewer, and records its own weekly price snapshots for 140 products: 481 dated observations taken by us since 2026-06-24. We research and we track prices. We do not conduct hands-on testing and we have not handled the products we recommend. Picks are chosen before commissions are considered, and no manufacturer can pay for placement.

What we do not do: test

We do not buy, borrow or lab-test products. Nobody at mid1 has handled the products in these guides. Every verdict on this site rests on three inputs, and each one has a limit worth knowing:

Published specifications and measurements
Key count, key action, velocity curve, pad sensitivity, aftertouch, polyphony, connector type, bus power draw. These are facts you can check against the manufacturer. What they cannot tell you is how a key bed actually feels under your fingers or how a pad grid holds up after a year of finger drumming, and figures like latency and battery hours are the manufacturer's own and are not comparable between brands. Where an independent source publishes measurements, we say so and name it.
Owner reviews read at volume
What we take from these is the pattern: the failure that keeps recurring, the feature nobody mentions, the complaint that only shows up after six months. What we do not take from them is the star average. Review counts track a product's age and shelf position, not its quality, so we never rank on them and never present a retailer rating as a verdict.
Our own weekly price record
This is the one input we generate ourselves. We snapshot the price of every tracked product once a week and keep the series, so a sale badge means cheaper than this product usually is, measured against prices we recorded and dated. The limit is depth: the weekly series starts on 2026-06-24, and the median product carries 2 checks over 14 days. We hold weeks of observations for most products, not years.

The rules

The Key-Count Rule

mid1 splits its keyboard guides by key count, because 25, 49, 61 and 88 keys are different instruments for different desks, and a "best MIDI keyboard" list that mixes them is comparing a laptop bag to a piano.

A 25-key controller is a travel and production tool that sacrifices range for a spot on a crowded desk. An 88-key weighted controller is trying to feel like a piano and is judged on action, not portability. Ranking them against each other tells you nothing useful, so every keyboard guide names its key count in the title and only compares controllers built for the same footprint.

The Three-Tier Rule

Every buying guide on mid1 names exactly three picks: a budget pick, a best-value pick and a premium pick, and the tiers are relative to that category rather than fixed dollar amounts.

A budget 25-key controller and a budget 88-key controller sit at very different prices. Tiers describe where a product sits inside its own category, so the budget pick is the cheapest thing we would still tell a friend to buy, the best-value pick is what most people should actually get, and the premium pick is the point where more money stops buying much more product. If a category has no honest trio, we do not publish a guide for it.

The Typical-Price Test

A discount only counts as a real sale on mid1 when the current price is below the median of our own tracked price snapshots, not when it is below the manufacturer list price.

List price is close to meaningless on most marketplace listings: items sit permanently at "20% off" a number nobody ever paid. We compare against the median of the prices we have actually recorded for that product, so a sale badge means the thing is genuinely cheaper than it usually is. Where Amazon shows its own strikethrough we display that figure attributed to Amazon, and it never becomes the typical price. A tracked low needs the price to have moved: we flag one only when we have recorded at least three different prices for that product and the current price is below every other price on our record, so a price that has never changed is never called a low.

The Weekly Check

mid1 runs its own price snapshot once a week and stores the result, so every price on the site carries the date it was observed.

The weekly series begins the week mid1 launched. Each run covers most of the tracked catalog rather than all of it, so a product sometimes carries the date of the run before. Once an observation falls more than ten days behind the latest run it loses its sale and tracked-low badges and is shown with its date instead, because a badge is only as fresh as the observation behind it. Snapshots are committed to the site as data rather than fetched at page load, so the history survives if an upstream source disappears. Every price on the site is a dated observation, not a live quote, and the retailer link always carries the true current number.

The Buy-Box Rule

A product only stays listed on mid1 while it has a live buy box from the retailer, so gear that quietly becomes third-party-only or unbuyable is pulled instead of left up.

Dead affiliate listings are the standard failure mode of recommendation sites. We treat the missing buy box, not the page returning an error, as the signal, because an unbuyable listing usually still returns a perfectly healthy page. Items that fail the check go to a replacement worklist rather than staying in a guide.

The Three-Strike Listing Check

mid1 only removes a product as dead after its listing fails three separate checks, because single failures are usually transient.

One failed request is noise: retailers rate-limit, and pages flap. Three confirmed failures is a removed product. This keeps the catalog from churning on temporary errors while still clearing genuinely dead listings.

No Paid Placement

No manufacturer can pay for a pick, a position or a rating on mid1; the only money the site makes is affiliate commission on links a reader chooses to click.

Commission rates differ between products, so the incentive to bias a ranking is real. The rule is simply that rankings are set before commissions are considered, and a pick is never changed because of what it pays. Affiliate disclosure appears on every page that carries a link.

One Named Reviewer

Every review, rating and verdict on mid1 is written by one named person, Elan Harpaz, rather than published under an anonymous house byline.

Ratings are given on a 1 to 10 scale and represent a single editorial verdict, not an average of many reviewers. Star ratings that come from a retailer listing are always labelled as the retailer's, shown separately, and deliberately kept out of our structured data.

What we do not do

  • We do not claim to have tested, tried or used a product. See above.
  • We do not accept payment for a pick, a rank or a rating.
  • We do not present retailer star ratings as our own score, or as consensus.
  • We do not publish a guide for a category where an honest budget, value and premium trio does not exist.
  • We do not quote a price without the date it was taken.

Corrections

If a spec, a price or a verdict on this site is wrong, tell us and we will fix the page and note the change on it. Reach the editor via the about page.

Related: our price-tracking data · who writes this