Keurig Kold withdrawn: what the public record can and cannot show

Suppose you are about to pay for tooling on a machine that runs on refill pods, and an investor asks whether customers will really pay that price. On 7 June 2016, Food Dive reported that Keurig was discontinuing the first generation of Kold, its at-home cold drinks system that had debuted in September (the library gives 2015), offering refunds and cutting about 130 jobs (Food Dive, 2016). A tempting reading is that the product did not fit how people behave. This post tests that reading against the report itself, then builds a test you can carry to your own product. The facts are summarised in the Keurig Kold case. My only source is Carolyn Heneghan's short piece for Food Dive, which I read on 1 November 2026.

What does the public report actually say about Kold?

Three things. Keurig is discontinuing the first generation of Kold and offering a full refund on the device. About 130 jobs go, 108 in Vermont, including people who made Kold pods and related support staff. A spokesperson said Keurig had learned a lot, including that consumers will embrace a system that delivers fresh-made cold drinks at home, and would build that into future systems (Food Dive, 2016). In the part Food Dive quotes, the spokesperson gives no reason for pulling the machine.

Everything else is commentary. Food Dive says Kold was "plagued from the start" by concerns about its suggested retail price, and that the US$299–369 tag, plus regular pod purchases, "always seemed out of touch" (Food Dive, 2016). That is the outlet's judgement, not a measurement. And 108 of 130 is about 83 per cent in Vermont; where the other 22 or so worked, it does not say.

Is there evidence that Kold misread user behaviour?

Not in this source. Behaviour is what people do: how many machines they bought, how often they used them, how many pods they reordered, how many they returned. The report gives none of those figures, and quotes no customer.

The two judgements it does contain are about different things. Keurig says people like the concept. Food Dive says the price was too high. Both can be true at once: people like the idea and will not pay that much for it. In my analysis that is one possible reading, not a conclusion. The spokesperson's line is a company statement, and "out of touch" was written the day the product was withdrawn. Neither replaces sales data. So the label "behaviour mismatch" has nothing to rest on here: the source neither supports it nor rules it out.

What do the four feasibility gates show for Kold?

This grouping is the post's own analysis, not the report's: the report has detail on the product and organisation gates, touches the industry gate lightly, and gives no figure at all on the finance gate. The library's four-part feasibility analysis asks about product, industry and market, organisation, and finance in that order. The post on checking feasibility before the business plan explains the framework.

The organisation row needs a careful read. Food Dive credits Kelley's support for Kold to a Burlington Free Press report. The idea that JAB leans towards coffee (Food Dive names Peet's, Caribou, Stumptown and the newly announced Krispy Kreme deal) and so Kold "didn't fit" is Food Dive's inference. Keurig did not confirm that this device is the last (Food Dive, 2016). March to the 7 June report is three months, but sequence does not prove cause.

Which five checks belong before you pay for a device with refills?

This part is mine, not the source's. Each check starts from something the report leaves exposed.

Priced against what. Name the cheapest thing your customers use now and its cost per use. The report gives Kold's price and names one direct competitor, SodaStream, but gives no price for it (Food Dive, 2016). So we cannot see the gap. You must.

Total cost of ownership. The phrase "plus regular pod purchases" shows the sticker price is only part of what the buyer weighs (Food Dive, 2016). The formula: device price divided by the number of uses in its life, plus the refill price per use.

Who champions it. If the project lives through one person, ask what happens when they go, and whether the product can still justify itself alone. For a small firm, "ownership change" may be a distributor, a lender or a large shareholder who shifts priorities.

Evidence first. Fix the threshold in advance. For example: 40 households put down a deposit at the intended price, and at least half reorder refills after eight weeks. Watch for the sunk-cost trap: once the tooling exists, people are tempted to lower the threshold to justify what is already spent.

Cost of withdrawal. In the report, the full refund and the job losses among pod staff are two costs that had already landed (Food Dive, 2016). In my analysis, a consumables line has its own people and its own stock, so it belongs in the project's numbers from day one.

Brand B, hypothetically: can a cold-tea machine beat a bottle of tea?

This example is hypothetical. Brand B does not exist and every figure is made up to show the arithmetic. B makes a countertop cold-tea machine that uses single-use pods. The machine costs VND 4,500,000 and each pod VND 12,000 per glass. A customer keeps the machine for two years and drinks one glass a day, so 730 glasses. The cheapest ready-made substitute is bottled tea from a shop at VND 10,000 a glass. Brewing loose tea at home is cheaper still, assume VND 2,000 a glass.

Cost per glass with the machine: 12,000 + 4,500,000 / 730 ≈ VND 18,164. That is VND 8,164, or 82 per cent, above the bottle, and more than nine times home brewing. Over two years: VND 13.26 million with the machine against VND 7.3 million with bottles.

If the customer drinks only 0.3 glasses a day, 219 in two years, the machine's share per glass rises to 4,500,000 / 219 ≈ VND 20,548, making VND 32,548 a glass, more than three times the bottle.

Option 2 brings the cost per glass close to bottled tea. But B loses VND 1,100,000 on each machine and recovers it only after 220 pods. If the average customer drops the machine after three months, about 90 glasses, B earns 90 × 5,000 = VND 450,000 and is still VND 650,000 down per machine. Across 1,000 machines, that is VND 650 million. B's payback period now depends on whether customers keep reordering pods, which is exactly what the Kold report does not tell us.

This fits if customers use the product steadily and the refill costs about what they already pay per use. It does not fit if demand is irregular, or if the cheapest substitute undercuts your refill from the first glass: no machine subsidy rescues that.

Going deeper: where does the four-gate framework come from, and how far does this report carry it?

The library records the four-part feasibility analysis as coming from Barringer and Ireland (2018). What I say about the framework rests on the library entry: four gates in order, each with a route to rework or drop the idea, and a warning that feasibility analysis answers "whether to go on", not the business plan's "how".

Three limits apply when it meets Kold.

First, the evidence is thin. The whole post rests on one short brief from a trade site. The judgements on price and ownership are the writer's, and the spokesperson's line is a company statement with no independent check. Food Dive takes the Kelley point from another newspaper and the SodaStream share move from another outlet; I report only what Food Dive says about them.

Second, hindsight. "Always seemed out of touch" was written on the day the product was withdrawn. In my analysis, knowing whether the price was out of touch would take first-months sales and return rates, which the brief does not give.

Third, a checklist does not weigh its own items. The framework says where to look, not which gate counts for more. For a machine with refills, I would tie the product gate and the finance gate together through total cost of ownership, because machine and refill sit on one bill for the customer. That is my reading, not the library's.