1.07Founder readiness and the minus-one stage

Tech-push vs market-pull, choosing the first problem

Stage: Minus-oneDeeptech

The core idea#

Most lab spinouts start with a capability looking for a problem. That is fine as a starting point, but the first product must solve a problem a specific buyer already pays to solve, and solve it so much better on the one metric they buy on that switching is worth their risk. This playbook shows how to list every possible application, score them, test the top three with real buyers, and commit to one beachhead while keeping a credible route to bigger markets.

Stage: minus-one to pre-seed · Lens: Deeptech · Reading time: about 30 minutes

Who this is for#

  • Researchers whose technology could serve many industries (new materials, sensors, platforms, enabling methods).
  • Founders who keep getting different advice about which market to pick.
  • Teams that have chosen a market but have no evidence buyers will switch.

What you leave with#

  • A scored list of 10–20 applications.
  • A 2x2 placing them by size of advantage and time to revenue.
  • A beachhead memo: buyer, buying metric, threshold, current alternative, budget owner and sales cycle.
  • A sequence of two or three follow-on markets that use the same core technology.
  • Kill criteria that tell you when to switch beachheads.

Definitions used in this playbook#

Term Meaning here
Tech-push Starting from what the technology can do and
looking for a use
Market-pull Starting from a problem buyers already spend
money on
Term Meaning here
Application A specific use of the technology for a specific type
of buyer
Buying metric The one number the buyer judges suppliers on:
cost per unit, yield, latency, assay time, weight,
lifetime
Threshold The level on the buying metric at which the buyer
would switch
Beachhead The first market you commit to: narrow enough to
win, and a base for the next markets

Why lab founders get this wrong#

The problem is rarely a lack of applications. A new material, sensor or method usually has dozens. The problem is choosing by the wrong criteria.

The tech-push trap#

Researchers are trained to value what is most impressive scientifically. Markets reward what is most useful to a buyer who already has a budget. These are often different applications. Five patterns show up again and again:

  1. Choosing the biggest market. "Batteries are a $100 billion market" says nothing about whether you can sell to anyone in it in the next two years. Big markets usually have entrenched suppliers, long qualification cycles and buyers who are slow to take risks.
  2. Choosing the most exciting application. The application that got you into Nature is often the hardest to commercialise, because it needs the most performance and the most new infrastructure.
  3. Picking on "better" rather than "better enough to switch". Being 20% better on a metric the buyer does not prioritise rarely wins a sale.
  4. Picking several markets at once. Spreading a team of four across three markets means three half-finished products and no reference customer.
  5. Ignoring the whole product. A buyer purchases a working solution, not a component. If your material needs a new manufacturing line, certification and an integrator before it is useful, the buyer is really being asked to buy all of that.

The economics of switching#

A buyer who switches to your technology takes on costs and risks that have nothing to do with your science:

  • Qualification: testing, certification and approvals, sometimes taking years in automotive, aerospace, medical or grid applications.
  • Integration: changes to their process, equipment, software or supply chain.
  • Supplier risk: you are a startup with no track record, and you may not exist in three years.
  • Career risk: the engineer who champions you and fails is blamed; the one who stays with the incumbent is not.

This is why practitioners often say a new technology needs to be several times better (sometimes called "10x") on the buying metric, or much cheaper, to win against an incumbent. The exact number varies by industry. The point is that your advantage has to cover the buyer's switching cost and risk, not just beat the alternative on paper.

What a good first problem looks like#

A good first application usually has most of these features:

  • The buyer has a pain they can put a number on, and an existing budget for it.
  • Your technology is far better on the one metric they buy on, not slightly better on many.
  • Qualification and regulatory hurdles are light, or you can meet them within 12–18 months.
  • You can deliver a complete, working solution, alone or with one or two partners.
  • There are a reachable number of buyers who talk to each other, so one reference sells the next.
  • Winning it teaches you something, and builds a capability, that the next market needs.

The application scoring method#

This method turns a long list of possible uses into a short list you can test. It takes a team about two weeks: one week to list and research, one to score and argue. The scores are not the answer; they make your assumptions visible so interviews can test them.

Step 1: List every application (aim for 10–20)#

For each, write one line: [technology] used by [buyer type] to [job], instead of [current solution] . If you can't name the buyer type and the current solution, it isn't an application yet.

Sources of ideas:

  • The introduction and discussion sections of your own papers, and of papers that cite you.
  • Patents that cite similar work (search Google Patents or Lens.org for your key terms).
  • Companies that have contacted your lab, asked for samples or attended your talks.
  • Adjacent industries with the same physical problem: separation, sensing, storage, catalysis, imaging.

Step 2: Name the buying metric and threshold for each#

For each application, write the one metric the buyer judges on, the current best value, your expected value, and a guess at the switching threshold. Mark each guess as "interview" (from a real conversation) or "desk" (from reading). Most will be desk guesses at first; interviews replace them.

Step 3: Score each application#

Score 1 (poor) to 5 (excellent) on seven criteria. The weights reflect what tends to decide early deeptech markets; adjust them for your field, but agree them before scoring.

Criterion Weight What a 5 looks like
Size of advantage on the buying
metric		 3 Several times better, or much cheaper, on
the metric the buyer cares most about
Pain with a budget 2 Buyer can put a number on the cost of the
problem and already spends on it
Buyer clarity and access 2 You can name 50 target buyers and reach
them through people you know
Time to first revenue 2 Paid pilot possible within 6–12 months
Low regulatory and qualification
burden		 1 No certification, or a short standard test
Low capital needed to deliver 1 Can make pilot volumes in or near the lab
Leverage to the next markets 1 Winning here builds capability and
references for larger markets

The maximum weighted score is 60.

Worked example (illustrative)#

A team has a new porous membrane that separates organic solvents from water far more selectively than polymer membranes. They list six applications and score them.

Application Advantage Pain Buyer Time Regulation Capital Leverage Total
(×3) (×2) (×2) (×2) (×1) (×1) (×1) /60
Lab filtration
consumables for
research labs		 4 2 5 5 5 5 2 48
Solvent recovery in
pharma plants		 4 4 4 4 3 4 4 47
Oil-water separation
in refineries		 3 3 2 2 3 3 3 32
Household water
purifiers		 2 3 2 3 3 2 2 29
Seawater
desalination		 3 4 2 1 2 1 3 29
Battery separators 3 4 2 1 1 1 4 29

How to read it. Two applications stand well clear. Lab consumables scores highest because it is fast, cheap and easy to sell, but the pain is low and it doesn't lead anywhere big. Pharma solvent recovery is nearly as high and has real pain and strong leverage into chemicals and other process industries. The three large "headline" markets (desalination, batteries, water purifiers) score lowest: they are huge, but the time to revenue, capital and qualification burden make them poor first markets.

The decision this sets up. Test pharma solvent recovery as the beachhead. Consider selling lab consumables in parallel as a low-effort source of early revenue and performance data, but only if it doesn't distract the team (see the beachhead section below).

Step 4: Argue about the scores#

Have each founder score independently, then compare. Large disagreements are the most useful output: they show where you need evidence. Write each disagreement as a question for interviews, for example "Do pharma plants measure solvent loss in money, and who owns that budget?"

Placing applications on a 2x2#

The total score hides the trade-off that matters most at minus-one: how big your edge is against how soon someone will pay. Plot each application on those two scores. The picture usually makes the choice obvious, and it is a good slide for advisors and investors.

Worked example · 6 applications by advantage and time to revenue

The example shows a common pattern. The famous markets (desalination, batteries) sit on the left: the technology has a real edge, but revenue is years away. The quick-revenue options sit at the bottom right or top right. The beachhead is the application in the topright quadrant with the most pain and the best route to the next markets, which here is pharma solvent recovery.

How to use each quadrant#

Quadrant What to do
Beachhead candidates (big edge, fast
revenue)		 Test the top two or three with interviews and
pilots. Pick one.
Strong edge, slow to revenue Your platform story and later markets. Keep
them in the pitch, out of the first 18 months'
plan. Look for a partner who could fund
qualification.
Quick revenue, thin edge Possible sources of early cash and data, if they
need little team time. Don't let them become
the company by accident.
Quadrant What to do
Avoid as a first market Drop them from the plan. Revisit only if the
technology changes.

Choosing the beachhead and sequencing the next markets#

The beachhead is the one market you commit to winning first. "Winning" means becoming the obvious supplier to a defined group of buyers, so that references and word of mouth do part of your selling.

The beachhead test#

Bill Aulet's Disciplined Entrepreneurship (MIT) turned beachhead selection into a repeatable step: segment the market, then pick one segment and ignore the others for now (O'Reilly preview). His criteria cover whether customers are well funded, whether your sales effort can reach them, and whether they have a compelling reason to buy (MIT OCW lecture slides). Summaries of the book also stress beating the competition, opening up future markets, and fit with the team's interests (Sunil Khandbahale), plus strong word of mouth between users in the segment (Rework).

Adapted for deeptech, ask seven questions of each top-quadrant candidate:

  1. Money. Do these buyers have budget, and do they already spend on this problem?
  2. Reach. Can your small team reach decision-makers directly, without a large sales force or a distributor?
  3. Reason to buy now. Is there a cost, regulation, deadline or failure that makes the problem urgent this year?
  4. Whole product. Can you deliver a complete working solution within 18 months, alone or with one or two named partners?
  5. Competition. Is there an incumbent who can match you quickly, or who controls the buyer relationship?
  6. Word of mouth. Do buyers in this segment talk to each other, attend the same events, and copy each other's suppliers?
  1. Leverage. Will winning here make the next market easier, through shared technology, manufacturing, certifications or references?

How big should a beachhead be?#

Smaller than founders expect. One widely cited guideline from Aulet's book is an annual addressable market in roughly the $20–100 million range for the beachhead itself: big enough to build a real business, small enough to dominate (Tian Pan's summary). A deeptech company's long-term market should be far bigger; the beachhead is only the way in.

Sequencing the next markets#

After choosing the beachhead, sketch the next two or three markets and what connects them. There are three common kinds of connection:

  • Same buyer, new product. Sell a second product to the same customers (for example, from solvent recovery to wastewater treatment in the same pharma plants).
  • Same product, new buyer. Sell the same product to an adjacent industry with the same problem (from pharma to fine chemicals to food processing).
  • Same core, new performance level. Use revenue and learning from the beachhead to reach the higher performance or lower cost a big market needs (from specialty separation to large-scale water treatment).

For the worked example, the sequence might be:

Stage Market What carries over
Beachhead (years
1–3)		 Solvent recovery in pharma
plants		 First manufacturing line, plant-level
reference customers, operating data
Next (years 3–5) Fine and specialty chemicals Same product with small changes;
references from pharma
Later (years 5+) Industrial wastewater, then
desalination		 Lower cost from scale; reliability data
from years of operation

Parallel side markets#

Sometimes a quick-revenue application can run alongside the beachhead, like the lab consumables in the example. This works only when it needs little engineering or sales time, uses the same production process, and produces data you need anyway. Give it a strict cap, for example no more than 10% of team time, and review it every quarter.

Testing your choice with evidence#

Scores are guesses until buyers confirm them. Test the top three applications in parallel for six to eight weeks, then commit to one. Playbook 1.19 covers how to run interviews; this section covers what to test and when to switch.

How many interviews per segment

Stage Interviews per segment What you should know by the end
First pass (weeks 1–3) 8–10 in each of three
segments		 Whether the pain is real and has a
budget; who the buyer is
Narrowing (weeks 4– 10–15 more in the top The buying metric and threshold; how
6) two purchases happen; sales cycle
Commitment (weeks 5–10 return visits in the Whether anyone will give a costly
7–8) top one signal: samples, data, pilot, LOI

That adds up to 40–60 interviews, consistent with the minus-one plan in playbook 1.04.

Questions that test each scoring criterion#

Criterion Ask Good sign
Size of advantage "What would a new solution have
to do on [metric] for you to
switch?"		 Their threshold is below what
you can deliver
Pain with a budget "What did this problem cost you
last year? Whose budget pays for
it?"		 A number, and a named
budget owner
Buyer access "Who else would need to agree to
try something new?"		 A short chain of approvals you
can reach
Time to revenue "How long did your last new-
supplier trial take, from first call to
purchase order?"		 Under 12 months
Regulation and
qualification		 "What tests would we have to
pass before you could use this?"		 Standard tests you can run or
pay for
Leverage "Who else in your industry, or in
other industries, has the same
problem?"		 Named adjacent segments

Kill criteria: when to switch beachhead#

Write these down before you start testing, so the evidence decides and not your attachment to the idea. Switch away from a segment if, after 15 interviews in it:

  • Fewer than a third of interviewees describe the problem as a top-three priority.
  • No one can name a budget that pays for it.
  • Your best technical result is still below the threshold most buyers name, with no clear path to close the gap in 12 months.
  • Every buyer says they would need a certification or qualification process longer than two years.
  • No one offers anything that costs them, even when asked directly.

Switching is not failure. It is the purpose of testing three segments at once. Record why each segment lost; the reasons are useful in the pitch ("we tested desalination and here is why it comes later").

Case files: how lab spinouts found their first market#

Both cases show a platform technology with many possible uses, an early market chosen for speed, and later markets that were much larger.

E Ink (MIT Media Lab, 1997): from shop signs to the Kindle

The technology. Microencapsulated electrophoretic display, or electronic ink: a paper-like display that holds its image with almost no power. MIT filed the first patent in October 1996 (Wikipedia), and the company was founded in 1997 by Prof. Joseph Jacobson, undergraduates J.D. Albert and Barrett Comiskey, and business founders Jerome Rubin and Russ Wilcox (Wikipedia).

The first market. The founders aimed at existing markets, starting with digital signs for shops and wristwatch displays (Channel Daily News). The first commercial product, in 1999, was an in-store advertising display for J.C. Penney (Salon). Retail signage was seen in the industry as the lowest-hanging fruit (MIT Technology Review).

The switch. By 2001, E Ink announced it would stop making its retail signs and focus on displays for handheld devices (MIT Technology Review), and it later put its effort into e- books (Channel Daily News). The first e-reader using E Ink was Sony's in 2004, and wide success came with the Amazon Kindle in 2007, about a decade after the concept was developed (Science Friday).

Lessons

  • The first market taught the team about manufacturing and reliability, but it wasn't where the value was. The buying metric that mattered most (readability with very low power) mattered most in handheld reading devices.
  • The big market needed an ecosystem: devices, content and pricing. Sony's first e- reader struggled for lack of affordable content (Channel Daily News). A great component is not enough until the whole product exists.
  • Today the technology also serves electronic shelf labels and signage again (Dealroom), so the early market was not wrong, only early.

Log9 Materials (IIT Roorkee, 2015): from cigarette filters to batteries

The technology. Graphene-based materials for filtration and energy. Akshay Singhal began graphene research as an undergraduate at IIT Roorkee and set up the company in 2015 (YourStory), later joined by Kartik Hajela (Wikipedia).

The first market. While working on air and water filtration, the founders found that their filter material worked on cigarette smoke. Their first product, PPuF, was a plug-on cigarette filter that removed up to half of toxins such as tar, priced at ₹30 and reusable for

five cigarettes (YourStory). Four months after incorporating in April 2015, they received ₹15 lakh of pre-seed funding from TIDES, IIT Roorkee's incubator (YourStory).

The next markets. They added an oil-spill sorbent product by 2018 (BuzzOnEarth), and moved into energy with graphene-based metal-air batteries (Digit). In 2019 Log9 raised a $3.5 million Series A led by Sequoia Capital and Exfinity Venture Partners (Wikipedia).

Lessons

  • A small, fast first product can bring in money, a patent and credibility. Log9's cigarette filter gave it its first granted patent (Digit).
  • The founders said graphene's slow adoption came partly from the high cost of making it and from researchers taking a sector-agnostic approach (Digit). Picking specific applications was their answer.
  • Moving from filters to energy was a large jump, effectively starting a new market. Plan for that cost when a beachhead doesn't lead naturally to the next market.

The pattern#

E Ink Log9 Materials
Core technology Low-power, paper-like
display		 Graphene-based materials
First market Retail signs (1999) Cigarette filters (2015–16)
Why it was chosen Existing market, quick to
sell into		 Quick to make and sell, discovered
by accident
Main later market E-readers Batteries and energy
What the first market
gave		 Manufacturing learning First patent, revenue, incubator
funding

Telling the platform story without losing focus#

Investors in deeptech want two things that sound contradictory: a narrow first market you can win, and a technology big enough to matter in huge markets later. The way to show both is a sequence, not a list.

The three-slide structure#

  1. The beachhead. One buyer, one problem, one number. "Pharma plants lose [X] a year to solvent waste. Our membrane recovers [Y]% more at [Z] cost. Three plants are piloting it."
  2. The sequence. The 2–3 next markets and what carries over to each: product, manufacturing, certifications, references. Use the table format from the beachhead section.
  3. The platform. The long-term markets (here desalination and batteries) with one line on what has to be true to reach them, and when.

Language that works and language that doesn't#

Instead of Say
"Our technology can be used in 12
industries"		 "We start in pharma solvent recovery, and the
same membrane serves fine chemicals and
water treatment next"
"A $50 billion market" "About 400 plants in our first segment, each
spending [X] a year on this problem" (bottom-
up, playbook 1.08)
"We're a platform company" "Each market we win lowers our cost and adds
references for the next"
"We're exploring several applications" "We tested three markets with 45 interviews
and chose this one because…"

What investors listen for#

  • That you chose, and can explain why the others lost.
  • That the beachhead's buyers have given costly signals.
  • That the later markets share your core technology, not just your ambition.
  • That the 18-month plan is about the beachhead only.

Mistakes and how to recover#

Mistake Why it hurts How to recover
Choosing the biggest
market first		 Long cycles, entrenched
rivals, capital you can't raise
yet		 Re-score with time to
revenue weighted at least as
heavily as size; look for a
smaller segment with the
same problem
Choosing the application
from your best paper		 Scientific novelty and buyer
value are different		 Interview buyers in three
segments before deciding; let
the threshold decide
Defining the buyer as an
industry ("pharma")		 You can't sell to an industry,
only to a role with a budget		 Name the role, plant type and
budget line: "plant heads at
mid-size API makers"
Competing on a metric the
buyer ranks third		 No one switches for it Ask every interviewee to
rank their buying metrics;
build for the top one
Running three markets at
once after choosing		 No market gets enough effort
to win		 Set a time cap on side
markets (for example 10%)
and review it quarterly
Ignoring the whole product Buyers can't use a
component alone		 List every piece the buyer
needs to use you; line up
partners for the gaps
Never revisiting the choice Markets and technology
change		 Re-score every quarter until
PMF; change only on
evidence
Changing beachhead
every month		 Signals no conviction to
investors and team		 Agree kill criteria in advance
and switch only when they
are met

Existing playbooks to read, and sources#

  • Bill Aulet, Disciplined Entrepreneurship (2nd edition, 2024). Steps 1–4 cover market segmentation, beachhead selection, the end-user profile and sizing the beachhead (O'Reilly). MIT OpenCourseWare also has Aulet's lecture on target customer
  • definition (MIT OCW).
  • Geoffrey Moore, Crossing the Chasm . Why early adopters and mainstream buyers want different things, and why a narrow first segment helps you cross from one to the other (Rework summary).
  • Steve Blank and Bob Dorf, The Startup Owner's Manual . The customer discovery method used to test each segment.

1.04 The minus-one stage · 1.08 Sizing a market that doesn't exist yet · 1.19 Customer discovery from the lab · 8.07 Technical risk vs market risk · 8.17 Build, buy or partner · 10.13 Platform or single product · 14.07 The deeptech pitch deck

Sources#

← 1.06 Founder visas: O-1, H-1B cap-exempt, startup visas1.08 Sizing a market that does not exist yet →