ARTICLE · Research

JTBD without the myths: how to find the customer's real job and how much to spend on it

Boris Kaptelov

The jobs-to-be-done method rests on a distinction that is easy to state and hard to apply: between what the customer asks for and the job they are actually trying to get done. Below — how to find that distinction in interviews and in data, what it turns into in product, price and packaging, and when such projects are worth taking on at all.

The easiest way to explain the method is by contradiction. Had Steve Jobs asked owners of push-button phones what they wanted, at best he would have heard about a lighter body and more comfortable keys. No one would have described a smartphone with a touchscreen. The same logic is ascribed to the age of carriages: ask a carriage owner what to improve and you will hear about the horse and the harness, not a car with an internal combustion engine.

Such leaps are the work of visionaries, and they happen rarely — only when a product truly breaks an industry open. That is a small share of all product innovation. In the vast majority of cases the opposite holds: the customer describes their job quite clearly, if you ask the right way. The whole method lives in that "if you ask the right way".

Drawing out a deep need without leading the customer toward your own hypothesis is a skill, not a questionnaire. The most common mistake, especially among founders in love with their product, is to stop probing the need and start selling the solution. For a team that wants to try it themselves, we recommend Rob Fitzpatrick's "The Mom Test" as a basic primer. But it is precisely the discipline of "do not coach the witness" that separates a real finding from flattering self-deception — and it comes with experience.

The customer knows the job — if you look the right way

The data source need not be an interview. In one of our projects we reconstructed customers' needs for a cosmetics retailer from tens of thousands of reviews on map services and social media. The result diverged from the client's hypotheses on both segments.

In the budget segment the company kept fighting on price. The reviews showed that war was already won: prices suited people, and the main pain was speed of service. Resources were going to something the customer had long stopped worrying about.

In the premium segment the client saw its job as improving interior quality and widening the assortment. The reviews said the opposite: the assortment already caused confusion, especially with Korean cosmetics, and the higher-spending shopper wanted not more choice but a smart helper to narrow it down: "take this and this". The work in this segment is curation, not shelf expansion.

Reconstruct the situation, don't ask for an opinion

A depth interview is valuable because it reconstructs context, not because it collects ratings. We ran a series of studies for a restaurateur opening a Five Guys-style burger chain in Moscow. That model has recognizable attributes: a blended patty and, above all, the way fries are served — an excess portion of good fries poured straight into the bag with the burger, plus a red-and-white design code.

Non-leading questions, sometimes with images shown, produced a non-obvious picture. Moscow consumers read the excess of fries as a plus. But chaotically dumping that excess reads not as generosity but as sloppiness. People are glad to get more tasty fries, but want them served neatly and in a larger container. The red-and-white tiling, in the Russian perception, evokes a cheap canteen, not a modern brand. The customer's job — generous, tasty, but clean and modern — called for pointed edits to the design reference while keeping the minimalism, not copying the original.

The job, not what you sell

Sometimes the real job lies nowhere near where the business looks for it. In a project for a children's sports club, geo-analytics and interviews showed that parents are not buying a sports section. The real job is a role model — one more senior adult in the child's life. The wording was blunt: "I want my son to become a man." Coaches were chosen not for athletic achievements but for the ability to build character and the drive not to give up. And people left most often over mundane logistics, not over the quality of the sport or the price. A club selling "sport" and "price" misses the job twice.

Hence a practical rule: your real competitor is often not the one you compare yourself to. The classic formulation came from Reed Hastings: in 2017 he noted that Netflix staff felt they competed with HBO, when in fact the service competes with sleep. The most underrated alternative to your product is to buy nothing, solve the job the usual workaround way, or go to bed earlier.

Time, precision and the cost of error

The first management decision in research is the trade-off between time and precision. You can be more precise and cut the risk of a failed launch, or you can try fast. The right answer is set by the size of the bet and the cost of error. A quality study of jobs takes, in our experience, one to three months for a mid-sized company. A small, reversible bet is reasonable to test live; a large, irreversible one — almost never.

Here two kinds of depth interview, often confused, diverge. A problem interview clarifies the job and context before you have a solution. A solution interview tests an already-formed hypothesis on a prototype. Skip the problem stage and the product manager guesses the market demand by intuition and merely checks how nicely they guessed.

The cost of error explains why industries behave differently. IT long did without deep research: you can fix an app in a week or two with a small team, so at first teams did not research at all, then moved to solution interviews. In manufacturing and FMCG it is the reverse: retooling a line for a new flavor is expensive, and the economics of a quality study add up on their own. Serial FMCG and retail companies made it their DNA — before putting a product on the shelf they always go into needs research to raise the share of successful launches. As the IT market matured, the cost of error rose there too, and product teams also went into research.

Hence the typical startup trap — a product "for oneself". The founder sees the need from their own experience and their circle's, and projects it onto a user with a different income, in a different region, in different circumstances. Often that turns out to be untrue, and the product misses the market.

From job to solution: product, price, packaging, channel

A discovered job is useless until it turns into decisions across four axes: product, price, packaging, channel. The working unit here is not a demographic segment but a demand space — a combination of occasion and need, functional and emotional. In Kantar's brand-growth methodology, the search for a new space is described precisely through new occasions and needs — through "why", not "who".

Price is dictated by perceived value on the specific job. The cosmetics case is about exactly this: in the budget segment no one intended to pay more for something that already suited them; the job was speed. Bain and Kantar record the same mechanic on the Chinese market: when the shopper sees no tangible benefit from the pricier option, they choose what is cheaper. A premium is justified only where the job demands a premium result.

Packaging and format are part of the job, not a wrapper. The burger case shows how the very same abundance sells or repels depending on the serving. And finally, growth more often lies in penetration than in premium alone: Bain, on Chinese-market data, shows that as brand penetration grows so do frequency and repeat purchases, forming a self-reinforcing cycle. JTBD helps pick the jobs and occasions that widen reach.

Assortment versus clusters

For a mature company the pain is usually the opposite of a startup's. A good product grows wide, and the temptation arises to close every micro-segment with a separate micro-launch. At some point assortment expansion must be stopped, and instead you gather users from interviews into clusters with a shared need — so one product and one message cover several sub-segments at once, cutting the number of SKUs, line retoolings and the costs that come with them.

The opposite extreme is a slide into customization on the principle "let the user assemble the options themselves". Anyone who has bought a premium car knows that not everyone finds sorting through options comfortable, and it often looks excessive. Though there is a segment for whom the choice itself is the pleasure. A simple analogy: some love a restaurant with a short menu where every dish is perfected and market-proven, others a wide menu that is interesting to explore. People differ, and the task is not to guess the "right" format but to surface needs, cluster segments, tie that to the company's strategy and carry it through to communication — to attract exactly those the product is made for.

Where to start

JTBD is useful exactly to the extent it reaches product, price, packaging and channel. So start not with interviews for their own sake but with two decisions: what is your bet and what is the cost of error. That determines whether you need a full study or a quick check will do.

If the launch is expensive or irreversible, it makes sense to run a JTBD diagnostic: reconstruct the jobs and demand spaces on your own customers — from interviews and from review data — separate real alternatives from imagined competitors, and translate all of it into decisions on product, price, packaging and channel.