For direct-to-consumer shipping lines running 30 to 120 containers per minute, servo-driven piston fillers and net-weigh gravimetric heads are the two workhorse platforms, with overflow and pump fillers filling specific niches [S1][S2].
Selection starts with three hard inputs: the product's viscosity band, the smallest and largest container with neck diameter, and the per-container legal-metrology tolerance, normally OIML R87 accuracy class XIII or XII and the EU 76/211/EEC three-packers rule for the relevant pack size [S1]. E-commerce SKUs typically span 5 ml sample sizes to 1 L family formats, so a single fill range rarely covers the catalogue and a dual-head or quick-change layout is common [S2][S3].
Filling Principles vs Viscosity Band
Gravity and overflow fillers stay cost-effective below roughly 100 mPa·s, while piston and lobe-pump fillers take over from 1,000 mPa·s upward, and servo-driven piston units are now specified as standard in personal care and food SKUs because they hold ±0.5% volumetric accuracy across 20 ml to 1,000 ml change parts [S2][S3][S4].
For powders and granules shipped DTC, auger fillers running 30 to 80 rpm and vibratory net-weigh heads are the two mainstream options, with auger suited to spice, supplement, and protein-powder SKUs at ±0.2 g repeatability and net-weigh preferred for free-flowing coffee beans, seeds, and snack mixes where density drifts with roast or moisture [S2][S4]. High-value or density-variable products point to net-weigh, with the target weight closed on a load cell before discharge to keep giveaway under 0.5% by mass [S1][S4].
Foaming, aerating, or shear-sensitive products, including many serums, probiotic drinks, and egg-white-based sauces, need snorkel or bottom-up diving nozzles and slower nozzle retraction to keep dissolved air out of the headspace; a piston filler without these features will routinely show 2 to 4% rejects on a cosmetic line [S2][S3].
Container, Change-Part, and SKU Strategy
Container geometry drives conveyor and star-wheel selection more than fill volume does: round bottles and jars run cleanly on rotary or inline conveyors, square or oval PET requires custom infeed scroll screws, and pouches need a dedicated form-fill-seal or pick-and-place architecture that a liquid filler cannot retrofit to [S2][S3]. E-commerce catalogues that mix 30 ml glass dropper bottles with 500 ml HDPE cylinders typically need two filler platforms under one PLC recipe store, not one wide-range head, because nozzle bore, diving depth, and drip-tray geometry all change with neck finish.
Change-part strategy is a procurement decision, not an afterthought: a quick-release nozzle set and servo-stored recipes cut SKU changeover from 25 to 40 minutes on a manual change to under 8 minutes on a servo-locked line, and the saving compounds across the 10 to 30 SKU switches a DTC line averages per shift [S1][S4]. For glass, EVA-coated change parts and gentle star-wheel infeeds prevent chipping at the 1.2 to 1.8 m/s conveyor speeds typical of personal-care lines [S3].
Compliance, Hygiene, and Explosive-Atmosphere Zoning
Legal-metrology compliance for prepackaged goods in the EU runs through the 76/211/EEC three-packers rule, which sets the average, the negative tolerated error, and the per-pack minimum, with the actual class selected by pack size under OIML R87; missing the rule costs the line its e-mark and the goods their free movement [S1]. For the US, NIST Handbook 44 paragraph 3.30 applies to similar effect on net-content audits.
Hygienic design for food, cosmetic, and pharma SKUs references 3-A Sanitary Standards and EHEDG for wetted surfaces, ISO 22716 for cosmetics GMP, and ISO 13408 with EU GMP Annex 1 for aseptic filling, with CIP capability and 316L stainless wetted parts as the default floor, not the upgrade [S1][S2]. Aqueous products with pH below 4 or above 9, plus solvent-based aerosols, push the line into ATEX 2014/34/EU zone 1 or zone 2 inside the filler enclosure, with the conveyor, level sensor, and drives all rated for the declared gas group [S1].
Conveyor and drive electronics on ATEX-rated lines must also satisfy the EU Machinery Directive 2006/42/EC for guarding, interlocks, and ergonomic operator access; the two directives are commonly specified together in the same procurement RFQ and verified during factory acceptance [S1].
Speed, Automation, and Throughput Math
Semi-automatic single-head bench fillers run 10 to 40 containers per minute and suit pilot or seasonal runs, while inline automatic fillers reach 60 to 120 cpm and rotary fillers push past 300 cpm on beverage SKUs, with the upper end reserved for high-SKU-count catalogues amortised across multiple SKUs [S1][S2][S4]. Modular systems that allow upgrades without full replacement are specified when a startup expects to scale from 30 cpm manual to 120 cpm automatic inside 18 months [S4].
Automation reduces labor cost on the filler station by roughly 30% and minimises operator error, with MES integration providing real-time OEE, giveaway, and reject-rate data on the same dashboard [S4]. For DTC lines, the per-container giveaway is the hidden cost driver: a 1 g overfill on a 50 g cosmetic jar shipped at 200,000 units per month is 200 kg of product per month, which exceeds the annual depreciation on a mid-range net-weigh head inside two years [S1][S4].
Integration points to verify before purchase: bottle unscrambler timing, rinser dwell, induction sealer power, and labeller placement downstream. Mismatched pitches between filler and labeller are the single most common cause of e-commerce line downtime, and a 5 mm pitch error between filler star-wheel and labeller star-wheel shows up as a 1 to 2% reject rate on the labeller, not the filler [S2][S3].
Comparison of Filler Types for E-Commerce Use
On four decision criteria relevant to DTC fulfillment, the four common platforms line up as follows. Gravity and overflow fillers are lowest cost and simplest to clean, but cap out below 100 mPa·s and 60 cpm on most lines. Servo piston fillers cover 1 to 50,000 mPa·s at ±0.5% volumetric accuracy, run 40 to 180 cpm, and change parts fastest, which makes them the workhorse for personal-care and food DTC. Net-weigh gravimetric heads hold ±0.1 to 0.2 g on powders and granules and are the only option that pays back on high-value SKUs with density drift, but cap at 60 to 80 cpm per head. Pump fillers handle shear-sensitive or abrasive liquids, integrate CIP cleanly, and run 30 to 120 cpm, but cost more than piston units at the same accuracy class [S2][S3][S4].
For multi-SKU e-commerce, the practical answer is rarely a single head. Most DTC lines run a primary servo-piston line for the dominant viscosity band, a secondary auger or net-weigh head for powder and granule SKUs, and a manual or semi-automatic bench filler for sample sizes and limited releases, all sharing one conveyor, one capper, and one coding machine for batch and date marking [S1][S4].
Selection Checklist and Common Failure Modes
The five factors repeated across procurement guides are: cost, efficiency, compatibility with product and container, ease of maintenance, and the maturity of the control platform, and skipping any of the five produces a measurable failure mode inside six months [S5].
Common failure modes in DTC installations: picking a wide-range head that is accurate at neither end of the range, ignoring nozzle retract speed on foaming products, undersizing the receiving tank so the filler starves at 80% of rated speed, and specifying a gravity filler for a viscous sauce to save 20% on capex. Each shows up as chronic giveaway, not as a hard fault, which is why it survives commissioning [S1][S2][S3]. On a related packaging-station decision, pharma distribution coding machine selection follows similar trade-offs between CIJ, TTO, and laser, since the coder sits directly downstream of the filler on most DTC lines.
Trackable signals for the next procurement cycle: servo-driven piston platforms replacing pneumatic pistons on retrofits below 100 cpm; CIP-ready stainless wetted parts as a default rather than an upgrade on cosmetic lines; and ATEX 2014/34/EU zone declarations now showing up in the same RFQ as the 76/211/EEC tolerance target, instead of being scoped to a separate hazardous-area line [S1][S4]. A secondary related signal is the convergence of coding and labelling on the filler hand-off: the same labeling machine platform that wraps a 50 ml jar will label a 1 L cylinder if the conveyor pitch is right, which pushes the bottle-handling decision back onto the filler spec, not the labeler spec [S1].
For the relevant spec sheets and selection criteria, see filling machine.