Product Box

The Product Box is a collaborative elicitation game invented by Luke Hohmann, in which participants, ideally the product's real customers or users, design its retail box as if it were on a shop shelf, then "sell" it to the rest of the room. By forcing them to choose what goes on the front, a name, a tagline, a handful of headline claims, and what they say out loud to close the sale, the game surfaces the features and, above all, the benefits customers value; the gap with the feature list the product team believes decisive is what the session brings to light. It belongs to the family of collaborative games, alongside the affinity map and the fishbowl.

Goal

The Product Box makes people express the value and benefit of a product through a commercial frame everyone already knows, the shop shelf: the customer no longer describes a function, they sell a product. In selling it, they reveal what, for them, justifies the purchase. The direct question, "what do you want?", by contrast produces a feature wish list, because a customer answers in the register you hand them, and product teams routinely confuse the features they are proud of with the ones customers care about.

The mechanism rests on an observation of Hohmann's: the customer wants to believe the product will solve their problem, so even when they write a feature on the box, they sell it by promising the benefit it delivers. The gap between the claims chosen for the front and the idealised list the team keeps in mind is where what Hohmann calls the gems of truth sit, the expectations the customer voices that the product does not yet meet. A session often surfaces a variant of this: a strength the product already has but leaves on the back.

The technique informs precise decisions: what to emphasise in the product and its messaging, which candidate features genuinely excite a buyer and which merely fill a comparison grid. It feeds prioritisation, product vision and positioning. Its deliverable is the boxes themselves, physical or digital, together with the transcribed sales pitches and the observers' notes. This material is photographed then analysed: the content is sorted into features, benefits and labels or slogans and into secondary themes, adoption, community, user experience, support, technology, price, value. The artifact of record is the box and its pitch.

Usage

When to use it

Framing a new offering: define the product and its messaging before the requirements of a launch are pinned.
Benefits in the customer's own words: hear which benefits the buyer plays up when they are the one selling the product.
Positioning and message unclear: let the market supply its own framing and vocabulary.
Real customers or users in the room: with a neutral facilitator, their buyer's instinct becomes the instrument.
A room locked in its internal assumptions: expose the gap between the dreamed product and what the buyer wants.

When not to use it

Feature set already fixed: the game surfaces value, it does not rank a list, prefer Buy a Feature or MoSCoW.
Representative or quantitative evidence needed: a small group of volunteers speaks for no population, run a survey.
Low psychological safety or a culture that reads the game as childish: the room will not engage, run structured interviews.

Description

The Product Box is a facilitated session that follows the arc common to collaborative games, an opening in which the box is built, an exploration in which it is sold and challenged, a closing in which the harvest is taken. What is its own lies in its artifact, the box, and in the mechanism it installs: by asking the participant to sell, it moves the talk off the ground of the feature and onto the ground of the benefit. The sequence runs as written, and each step has its characteristic mistake.

Players and observers

The participants are, as far as possible, real customers or users: the value of the game comes from their consumer instinct, and letting the product team play in their place is the first way to waste it. Insiders reconstruct the internal feature list, and the game then confirms the assumptions it was meant to test. You work in small groups of three to six, and you run several groups in parallel, because it is the variation between the boxes that is instructive.

The facilitator is neutral. They explain the rules, keep time and, above all, keep the debrief centred on benefits. Their neutrality is a load-bearing condition: to every "are we allowed to …?", the answer is yes, without exception. Any reservation, any steer on the content, contaminates the customer's own view and drags the game back toward what the facilitator expected. At least two observers complete the set-up, with distinct roles: one follows the seller and notes which benefits they push, the other follows the room and notes who nods, who questions, who stays unmoved. The audience's reactions are often the richest material, and they escape anyone who watches only the box.

The materials

The materials are deliberately poor, and that is part of the game. Blank boxes the size of a cereal box, markers, coloured paper, pre-cut stickers of the "New" or "Exclusive" kind, scissors, glue, tape and old magazines to cut from. On each table, two or three real retail boxes provide a visual reference, and a sheet of butcher paper protects the surface and invites writing. Buying and laying out these supplies is part of the preparation. On timing, allow thirty to forty-five minutes to build a box, five to ten minutes per box to sell it, so a session of an hour and a half to two and a half hours depending on the number of groups. A timer or a gong for the selling round adds a playful spring and protects the closing.

Running the session

Frame and warm up
The facilitator shows the reference boxes and has the group name what a real box carries: a brand and a product name, a tagline, a few headline claims on the front, the feature list and the small print on the back, a price flash, badges and stickers. This survey gives the group a shared visual grammar before it sets off.
Build
Each group designs the box of the product its members would buy. The practical convention, and a good teaching aid, puts the name, the tagline and the few claims worth shouting on the front and moves the fuller list of features and details to the back. This front-back split is a convenience: Hohmann treats the box as a single three-dimensional canvas and mandates no division of the faces. Present it as a shaping tool. Throughout this phase, the facilitator answers yes to every request for permission and feeds in no content.
Sell
A member of the group stands, the facilitators sit, and they pitch the box to the room as if closing a sale. This is the heart of the technique, and the mistake that ruins it is to read the features written on the box and stop there. The value is in the benefits the seller voices out loud, in the way they turn "generates the AVS statements" into "you no longer spend your Sundays on payroll". A facilitator who photographs the boxes and skips the pitch and the observation loses the essence of what the game produces.
Debrief and harvest
Photograph every box, which the participants keep, transcribe the pitches, then sort the content: features, benefits, labels, slogans and the secondary themes already named. Then look for the gap between the idealised product the team described and what the boxes actually talk about. This harvest is where the analysis is manufactured: sorting the transcript is what makes the session an elicitation. It happens while the group is still there and remembers what a word meant.
Vote, optionally
Electing the best box, your own excluded, is a good closer and a way to award attention. It is not a prioritisation result. The winning box measures the room's engagement, it is not the backlog, and reading it as "the requirements" is a misreading: to rank a known list, the exit is Buy a Feature or MoSCoW.

AI considerations

Upstream, AI is useful on the shaping of the preparation. A language model drafts the facilitation script, draws up the shopping list of supplies, prepares the briefing on the reference boxes and pre-writes labels and taglines calibrated to the domain, all shaping work where it is fast. The goal of the session, though, stays decided by a human who knows the stakes.

Downstream, it changes the economics of the harvest. Automatic speech recognition transcribes the recorded pitches, then a model groups the transcribed claims into Hohmann's categories, features, benefits, labels and into the secondary themes, which speeds up the sorting step, the most tedious one. A vision model helps index the box photos, and a first-pass positioning summary can be produced for the team to challenge.

Three limits are hard. AI cannot play the game: the value rests on the consumer instinct of real customers and their lived experience, and a "generated" box is only the internal-assumption trap in a new costume, since it invents plausible claims, exactly the fiction the technique exists to break. Reading the room escapes the machine in the same way: the richest signal is who nods, who stiffens, what goes unsaid during the pitch, and that takes an observer in the room. Finally, the nuance between a written feature and the benefit actually sold is a matter of judgement, and the recordings and photos of customers are personal data whose consent and retention fall to the facilitator.

Examples

A Swiss SME publishes Salaris, a payroll and human-resources SaaS. In a workshop, a group of customers designs the box of the product they would buy and sells it to the room.

Put forward by the customer Filed on the back by the vendor

Front

Salaris New

Swiss payroll done in three clicks

AVS/AI/APG and LPP statements generated on their own
Salary certificate ready for taxCustomer
Swissdec certifiedCustomer
Withholding tax handled, every canton

From CHF 39.-/month

Back

On the back, in small print

Payroll engine, tax scales and contributions kept currentVendor
Direct export to the AVS compensation fund
Quarterly VAT statement
Family allowances handled automatically
Multi-canton, three national languages
e-banking reconciliation and TWINT payments
Support in French, German and Italian

The Salaris box built by a customer group: the front shouts the benefits the customer chose, the back lists the features. Swissdec certification and the tax-ready salary certificate moved to the front, where the vendor had filed them on the back.

The lesson is a single reordering. The group put "Swissdec certified" and "salary certificate ready for tax" on the front, where the vendor had filed both among the back-of-box details. What the vendor foregrounds, its payroll engine, stays on the back and is never once named in the pitch; what the customer shouts is the peace of mind of being compliant and a document ready for tax. That reordering is the elicited requirement. Everything else on the box is there only to make it legible.

Cost

Phase	Level	Justification
Preparation	Medium	Recruit real customers, book a room, buy and lay out the craft materials, gather reference boxes and brief the observers on what they are to record.
Execution	Low	One facilitated session of about two hours is enough, with cheap materials and no tooling to install.
Documentation	Medium	Photograph the boxes, transcribe the pitches and sort the content into features, benefits and themes: the harvest is where the effort lands.

Tooling

The default tooling is physical: cardboard boxes, markers, coloured paper, stickers, scissors, glue, old magazines, a timer or a gong, plus a camera or a phone to photograph the boxes and record the pitches. Nothing to provision, nothing to license: the tactile handling belongs to the mechanism, and impoverishing it weakens the game.

Remotely, collaborative whiteboards (Miro, Mural, FigJam) carry a workable variant: a box template with frames for the front and the back that the group fills in together. You lose the tactility and part of the shared energy, you keep the framing, the constraint of choosing what goes on the front and the exercise of the pitch. This is the fallback when the team is dispersed and no in-person workshop is possible.

For the analysis, any affinity-sorting tool suits the grouping of the harvested claims, a wall of sticky notes or its digital equivalent, and the affinity map is its natural continuation. Transcription of the recorded pitches completes the set-up on the text side.

Sources

Hohmann, L., Innovation Games: Creating Breakthrough Products Through Collaborative Play (Addison-Wesley Professional, 2006): the Product Box is Luke Hohmann's method and this book is its original source.
Innovation Games, Product Box: the game presented by its author.
IIBA, A Guide to the Business Analysis Body of Knowledge (BABOK Guide) v3, §10.10 Collaborative Games: frames the family of collaborative games and names the Product Box, descriptively.