We benchmarked 16 cordless stick vacuum models — here is what separates a reorder from a return

What we tested

The full published cordless line: 16 models, suction 9–34 kPa, motors 100–500 W, 11.1–40 V platforms — every figure taken from the working-specification ledger, the same numbers printed on each model page. We looked at four dimensions a private-label buyer actually earns money on: suction relative to input power, the voltage platform behind the runtime claim, the certification family the model travels with, and the container loading that decides landed cost. One question ran through the benchmark: which column on a quotation predicts a reorder?

The benchmark table

Model	Power	Suction	Voltage	Units / 40HQ	File family
HJ 100B	100 W	9 kPa	11.1 V	200	GS family
HJ 100BL	100 W	12 kPa	11.1 V	200	GS family
HJ 150D	150 W	18 kPa	—	200	GS family
HJ 207	150 W	12 kPa	22.2 V	200	SDS family
HJ 208 Pro	250 W	24 kPa	22.2 V	200	SDS family
HJ 209 Pro	350 W	30 kPa	25.9 V	200	SDS family
HJ 209 Ultra	500 W	34 kPa	29.6 V	200	SDS family
HJ 1901	150 W	9 kPa	21.6 V	2,680	GS family
HJ S25 sweeper	100 W	—	40 V	8,900	GS family

All figures from the Cliia 49-model specification ledger (2026); eight further variants (folding-tube and display configurations of the above) share their parent platform's figures and are omitted for readability. Full 16-model table: cordless stick vacuum line.

Finding 1 — Wattage is a cost label, not a performance label

The table's quietest row is its loudest finding: the two 100 W builds in the line deliver 9 kPa and 12 kPa respectively — a 33% suction gap at identical input power. Meanwhile the step from 250 W to 350 W (24→30 kPa) and 350 W to 500 W (30→34 kPa) shows diminishing returns at the top. Procurement instinct reads watts as performance; the data reads suction-per-watt and the motor/airflow build as the real tier boundary. The practical translation for buyers: never accept a quotation that quotes watts without kPa beside it — the pair is the specification, and every model page in the line publishes both precisely so the pair can be checked.

Finding 2 — Two file families, five markets, one review

The second finding is structural: the line splits into a GS-carrying family (CE / GS / EMC / RoHS / CB / UN 38.3) and an SDS-documented family (CE / EMC / RoHS / CB / UN 38.3 / SDS), and both cover EU, GCC & Levant, Latin America, ASEAN and North America. A distributor can therefore build a three-tier ladder — entry 100 W, mid 150–250 W, flagship 350–500 W such as the HJ 209 Ultra — entirely inside one compliance review, provided the family is chosen to match the market's expectation (GS where German-style assurance sells; the SDS-documented family where the battery shipment file leads). Confirm which family your quotation names before comparing prices across suppliers; the quality and compliance dossier shows what each family contains.

What this means for your first order

The loading column is the sleeper: fifteen of sixteen models load a flat 200 units per 40HQ by design, so a three-tier ladder costs nothing in container terms — entry, mid and flagship ride together. The exceptions are where the mixing leverage lives: the compact sweeper at 8,900 units per 40HQ is the cube-filler that lets a floor-washer-heavy order top off without a second container, and the 2,680-unit model fills the middle. Compose the ladder first (three tiers, one file family), then use the high-cube models to close the container.

How to verify these specs before you pay

• Ask for the model's working-spec sheet with wattage and kPa on the same page — one without the other is a red flag
• Match the quotation's certificate family (GS-carrying vs SDS-documented) to your market's expectation before comparing prices
• Require UN 38.3 summaries in the quotation for any lithium platform — they are shipping documents, not brochures
• Request the per-model loading plan for your mixed 40HQ; a factory that publishes loading counts per model has already done this maths
• Run a 30-day sample on your market's floor types before the production order — the tier decision survives contact with real carpets or it doesn't