NPD as a Service: Bench Sample to First Export Run

Buyer’s snapshot

In 2025, 83% of food and beverage brands raised their new product development budgets, up from 76% in 2024 (TraceGains, 2025 NPD Report, n=190).
Published estimates of first-year failure for new packaged-food products still sit between 70% and 90%. Most of that outcome is settled long before the first pallet ships.
Three gates decide it: the bench sample, the pilot batch, and the first certified export run.
Silk Foods Ceylon runs development and production on one cellular manufacturing floor under BRCGS and FSSC 22000 V6, so a pilot batch does not trigger a separate audit cycle.
The stage table below is the part to forward to your operations lead.

Most contract manufacturers in South Asia will quote you a price per unit. A smaller number will develop the formulation first, then quote. A handful will do both inside the same certification scope, on the same floor, with the same quality team signing off at each gate. That last group is where R&D and NPD stop being a favour and start being a service with a schedule attached. This piece walks the three gates in order, states the weeks each one typically takes, and separates the decisions the buyer owns from the decisions the manufacturer owns.

What does R&D and NPD as a service actually mean?

R&D and NPD as a service means the manufacturer carries the formulation work, the pilot trials, and the specification build, and the buyer carries the commercial calls. In 2025, TraceGains surveyed 190 food and beverage quality, R&D, and regulatory professionals for its NPD Report and found 83% planning higher development budgets, up from 76% the year before. Budget is not the constraint. Sequencing is.

The distinction matters because it changes who absorbs the cost of being wrong. When a brand owner develops a formulation independently and then hunts for a co-packer, every plant-side constraint discovered late becomes a reformulation. When the development happens on the plant that will eventually run it, the constraint shows up at the bench, where it costs a week instead of a quarter. Silk Route Ventures (SRV) structures the work this way for a straightforward reason: the R&D bench and the production line at Silk Foods Ceylon (SFC) sit in the same 10,000 sq ft facility in Matale.

That structural point is covered in more depth in R&D and NPD at Silk Foods Ceylon, which maps the capability itself. This post is about the gates.

Gate 1: the bench sample, and what it can and cannot tell you

A bench sample is a small hand-made trial, usually a few hundred grams or a few units, produced to prove that the concept holds together at all. It answers taste, colour, texture, and rough ingredient cost. It does not answer yield, throughput, or shelf life. Buyers who treat a good bench sample as a green light for a purchase order are the buyers who get surprised at gate two.

What the bench stage is actually for is narrowing. A brief that arrives with four flavour directions leaves the bench with one or two. A brief that arrives with an unrealistic clean-label constraint leaves with the constraint priced. At SFC the bench cycle typically runs two to four weeks per iteration, and most projects take two iterations. Samples then ship door to door by international courier at three to five business days transit.

Where the NPD route doesn’t fit Brands that want a formulation developed at no cost against a vague volume promise. Brands that need a finished retail SKU inside eight weeks. Brands whose target landed cost is below the certified-origin band, where an uncertified regional co-packer is the more honest answer. The development route is built for buyers who intend to own a specification, not for buyers shopping a sample.

Gate 2: the pilot batch, where the formulation meets the plant

The pilot batch is where a formulation stops being a recipe and becomes a process. It runs on production equipment at reduced volume, which is the only way to measure real yield, real cycle time, real fill weights, and the way an ingredient behaves under actual shear, heat, or vacuum. Food Safety Magazine has argued this point for years: the benchtop-to-scale-up transition is where most food-safety and quality assumptions get tested for the first time.

Two things typically change at this gate. The first is cost, because bench yields flatter the real number. The second is the specification itself, because a pilot exposes tolerances the bench never had to hold. A vegan spread that sets beautifully in a 200 g bench trial may separate in a 3,000 jars per day fill, and the fix is usually a hydrocolloid ratio, not a new product.

Capacity is what makes a pilot cheap or expensive to run. SFC’s semi-liquid and spread line runs 3,000 jars per day, the beverage line 2,500 bottles per day, patties 15,000 units per day, nuggets 30,000 units per day, and the spray-drying line 50 kg per day. A pilot at 5% of a line’s daily capacity is a scheduling question. A pilot that needs a dedicated line build is a project.

SRV finding: across the SRV development desk, the single most common cause of a pilot re-run is not the formulation. It is packaging chosen before the process was fixed. A jar geometry, a film specification, or a closure picked at concept stage frequently fails the fill or the seal integrity test, and the brand then reworks the pack after having already paid for artwork.

That pattern shows up often enough that the SRV team now asks for target pack format at the first call, not the third. Pack decisions made at gate one cost a conversation. The same decision made after a pilot costs an artwork revision, a new tooling lead time, and usually four to six weeks.

Gate 3: the first certified export run

The first commercial run is the gate where certification, documentation, and freight stop being background and become the schedule. Production lead time from purchase order to dispatch at SFC is two to three weeks. Sea freight adds three to four weeks to the EU and Australia, four to five weeks to the United States. Those are the numbers to build a launch date on, not the production number alone.

The certification position is what removes a step here rather than adding one. SFC holds BRCGS and FSSC 22000 V6 across its processing scope, with USDA Organic and EU Organic on the relevant SKUs. Because the facility runs a cellular manufacturing layout across 400 plus products, a new SKU introduced through the development route sits inside the existing audit scope rather than triggering a fresh certification exercise. For a buyer whose retail listing depends on a GFSI-recognised standard, that is the difference between a launch quarter and a launch year.

Every dispatched order carries the same document pack: commercial invoice, packing list, bill of lading or air waybill, certificate of origin, phytosanitary certificate, an organic transaction certificate where the SKU is certified, and a batch certificate of analysis. Buyers unfamiliar with reading the last of those will find per-lot heavy metal and pesticide testing explained here.

How long does each gate take?

Total elapsed time from a first brief to a first certified export run typically runs 16 to 30 weeks for a format SFC already produces, and longer where a new process or a novel ingredient is involved. The table below is the planning baseline, not a quotation. Iteration count, pack sourcing, and shelf-life confirmation are the three variables that move it most.

Gate	What the buyer brings	What the plant produces	Typical elapsed time	The decision at the gate
Brief and scoping	Target market, claim platform, pack format, target landed cost, volume forecast	Feasibility read, ingredient sourcing options, indicative cost band	1 to 2 weeks	Is the target cost achievable inside the claim platform?
Bench sample	Sensory direction, competitor benchmarks, label constraints	Hand-made trials, ingredient declaration draft, revised cost	2 to 4 weeks per iteration, usually 2 iterations	Which single direction goes to pilot?
Pilot batch	Approved bench direction, confirmed pack specification	Line trial at reduced volume, yield and cycle-time data, process parameters, specification draft	3 to 6 weeks including scheduling	Does the real yield support the target cost?
Shelf-life confirmation	Claim duration required by the destination retailer	Accelerated study, parallel real-time study, stability data	6 to 12 weeks accelerated, running in parallel with pack finalisation	Is the date code defensible to the retailer?
First certified export run	Purchase order, artwork, destination compliance sign-off	Production run, batch COA, full export document pack	2 to 3 weeks production, plus 3 to 5 weeks sea freight	Ship, or hold for a regulatory dependency?

Shelf life is the stage buyers most often try to compress. Accelerated testing can shorten a twelve-month question into a matter of weeks by holding product at raised temperature, but the Institute of Food Technologists is direct about the limit: the more extreme the acceleration, the less the result resembles real-world performance, and the recommended practice is to run a parallel real-time study alongside it. An accelerated number opens a launch. A real-time number defends it during a retailer audit.

Which decisions does the buyer own?

The buyer owns every decision that is commercial, and the manufacturer owns every decision that is technical. That line sounds obvious and is violated constantly, usually in the direction of a brand asking a plant to pick a flavour direction, or a plant quietly setting a target cost by choosing ingredients.

The buyer owns: target market and its regulatory framework, claim platform and label copy, pack format and artwork, target landed cost, launch volume forecast, brand-side sensory preference, and the go or no-go at each gate.

The manufacturer owns: ingredient specification and sourcing, process parameters, food-safety controls and HACCP, yield and throughput, batch documentation, and the honest statement of what a line can and cannot hold.

Two areas need writing down before gate two rather than after. The first is formulation ownership. SRV’s standard position is that a formulation developed to a buyer’s brief and funded through the project belongs to that buyer, with the underlying process know-how remaining the manufacturer’s. Get whichever version applies stated in writing early, because it becomes contentious only when a relationship is ending. The second is payment structure: orders under $10,000 run 100% in advance by bank transfer, and orders of $10,000 or above run 50% in advance with the balance against scanned shipping documents. SRV does not extend deferred-payment terms, and knowing that at gate one avoids a renegotiation at gate three.

Buyers running this evaluation for the first time will find the supplier-side questions worth asking collected in qualifying a coconut and plant-based contract manufacturer.

Where the regulatory clock, not the plant, sets the timeline

Some launches are not limited by the manufacturer at all. If a formulation contains an ingredient without a significant history of consumption in the European Union before 15 May 1997, it falls under the EU novel food framework, and the statutory path is long: the European Food Safety Authority has nine months to adopt an opinion on a valid application, and the European Commission then has seven months to put a draft implementing act to the Standing Committee. Published analyses of actual cases put the practical average well beyond that, closer to two and a half years.

The United States sets a different kind of clock. Any foreign facility manufacturing, processing, packing, or holding food for US consumption must be registered with the Food and Drug Administration, must name a US agent, and must renew that registration in the October to December window of each even-numbered year. Prior notice is filed before every individual shipment. None of that is difficult, and all of it is disqualifying if it is discovered late.

The practical instruction is simple. Raise the regulatory question at gate one, in the same conversation as target cost. A novel ingredient discovered at the bench is a formulation choice. The same ingredient discovered after a pilot is a two-year delay wearing a formulation’s clothes. Buyers building for both the EU and US markets should also read the plant-based and functional food contract manufacturing overview for the format-level view.

Frequently asked questions

How long does it take to go from a bench sample to a first commercial run?

For a format Silk Foods Ceylon already produces, 16 to 30 weeks is the realistic planning band: two to four weeks per bench iteration, three to six weeks for a pilot, six to twelve weeks for accelerated shelf life running in parallel, then two to three weeks of production and three to five weeks of sea freight.

What is the difference between a bench sample and a pilot batch?

A bench sample is hand-made at a few hundred grams and answers sensory, appearance, and rough cost. A pilot batch runs on production equipment at reduced volume and answers yield, cycle time, fill accuracy, and process stability. Food Safety Magazine treats this benchtop-to-scale-up transition as the point where quality assumptions are first properly tested.

Can Silk Route Ventures develop a custom formulation for a US or EU brand?

Yes. The SRV R&D and NPD team scopes formulations in-house at the Matale facility and runs pilots on the same lines that will produce commercially, under BRCGS and FSSC 22000 V6. First-run MOQs vary by format: 1,500 jars for spreads, 1,250 bottles for beverages, 5,000 to 10,000 units for frozen patties or nuggets.

Who owns the formulation developed during an NPD project?

SRV’s standard position is that a formulation developed to a buyer’s brief and funded through the project belongs to that buyer, while the underlying process know-how stays with the manufacturer. Because practice varies across the industry, buyers should have the ownership position written into the development agreement before the pilot stage.

How Silk Route Ventures can help

Silk Route Ventures contract-manufactures finished CPG SKUs (spreads, sauces, and semi-liquids in glass jars at 3,000 jars per day; vegan patties at 15,000 units per day; vegan nuggets at 30,000 units per day; spray-dried plant milks; jackfruit-in-brine for plant-based meat formats; functional beverages at 2,500 bottles per day) under one roof at the Silk Foods Ceylon facility in Matale. The site holds BRCGS and FSSC 22000 V6 across its processing scope, with USDA Organic and EU Organic on the relevant SKUs. The cellular manufacturing layout means a new SKU introduced through the development route does not require a separate audit cycle, and the R&D and NPD team scopes the formulation in parallel with the production plan. Contact us to send an inquiry for a co-manufacturing capability briefing tailored to your SKU and target launch volume.

Sources

TraceGains, “2025 NPD Report” (survey of 190 food and beverage professionals), reported by Food Industry Executive, 2025. Retrieved 2026-08-20. https://foodindustryexecutive.com/2025/08/fb-brands-double-down-on-new-product-innovation-but-many-lack-digital-infrastructure/
European Commission, “Novel food authorisations,” Regulation (EU) 2015/2283. Retrieved 2026-08-20. https://food.ec.europa.eu/food-safety/novel-food/authorisations_en
US Food and Drug Administration, “Online Registration of Food Facilities.” Retrieved 2026-08-20. https://www.fda.gov/food/registration-food-facilities-and-other-submissions/online-registration-food-facilities
Institute of Food Technologists, “How to Fast Track Your Shelf Life Testing,” Food Technology Magazine. Retrieved 2026-08-20. https://www.ift.org/food-technology-magazine/safety-and-quality-how-to-fast-track-your-shelf-life-testing
Food Safety Magazine, “From Benchtop to Scale-Up: Food Safety Considerations for New Product Development.” Retrieved 2026-08-20. https://www.food-safety.com/articles/9243-from-benchtop-to-scale-up-food-safety-considerations-for-new-product-development
Sri Lanka Export Development Board, “Sri Lanka’s Export Performance Exceeded US$ 17.2 Billion in 2025.” Retrieved 2026-08-20. https://www.srilankabusiness.com/news/sri-lankas-export-performance-exceeded-us-17.2-billion-in-2025.html

NPD as a Service: From Bench Sample to First Certified Export Run