Category Design for Startups: Create a New Category or Win an Existing One?

  1. Blog
  2. Strategy & Positioning

By Saroj Jha · October 5, 2026 · 12 min read

Create a new market category only when the label buyers already use makes them judge your product on the wrong test, and you can afford to teach them a new one. Most early-stage startups do better entering a category buyers know, with a sharp difference. The five steps below show how to decide, with one example company.

In three numbers

  • $7.1 billion CRM applications market in 2002, per IDC: the budget Salesforce's new "on-demand" label was sold into[1]
  • 1.5 million visits a month to HubSpot's blog in 2014, as it taught buyers what "inbound" meant[2]
  • 16 years of US satellite radio, in which attention moved from the category to the firms inside it once it was accepted[3]

Category design means deliberately choosing, or creating, the label buyers use to file your product: the words they search for, the shelf they put you on and the budget they pay you from. It is a popular idea, and a costly one when it goes wrong. A new label can make you stand out. It can also make you invisible to the people who were about to buy.

Research on market categories explains both outcomes. Studying US public companies from 1985 to 1994, Ezra Zuckerman found that "the stock price of an American firm was discounted to the extent that the firm was not covered by the securities analysts who specialized in its industries"[4]. His argument: when the experts who judge a category cannot place you, demand suffers. Buyers work the same way. So the useful question is not "can we invent a category?" but "which label gets us found, compared fairly and paid?"

Running case: Example Co. Fictional

Example Co. is a seed-stage MedTech startup with a benchtop blood analyzer for outpatient clinics. From a fingerstick sample, it reports routine tests such as HbA1c (a common blood sugar test) and a lipid panel in about 12 minutes, so the clinician can discuss the results during the visit instead of waiting a day or two for an outside lab. In the case it is FDA-cleared and CLIA-waived (a US category for simple lab tests). Its first customers are independent primary-care practices with 3 to 10 clinicians. The founders keep debating one question: is this a new category, "in-visit diagnostics", or a better kind of "point-of-care testing"? Example Co. and all its numbers are illustrative.

Step 1: Write down the label your buyers already use

Before you name anything, find out how buyers file the problem today. Listen to the words they use on calls, the terms they type into search, the budget line the money comes from and the products they compare you with. That set of words is your current category, whether you chose it or not.

A category does three jobs for a buyer. It helps them find options, compare those options on agreed tests and pay from a budget that already exists. A product that fits no familiar label loses all three at once.

In medical products, some labels come from regulators. "CLIA-waived" is one. The FDA, quoting the CLIA statute, describes waived tests as "simple laboratory examinations and procedures that have an insignificant risk of an erroneous result"[5], and a waived test can be run by laboratories with a Certificate of Waiver[6].

Example Co., step 1 Fictional

In 12 calls with practice managers and physician-owners (fictional), the words that come up are "point-of-care testing", "in-house lab" and "CLIA-waived". Nobody says "in-visit diagnostics". Practices pay for these tests from the same budget line as their other lab supplies, and they compare products on how long results take, which tests the analyzer runs, cost per test and how much staff time it needs.

Figure 1 A category label tells buyers where to look, what to compare and who pays

Words heard in 12 calls with practice managers and physician-owners

  • "point-of-care testing"
  • "in-house lab"
  • "CLIA-waived"
  • "in-visit diagnostics"0 of 12 calls
The label "point-of-care testing" The words buyers search for and say
  1. Find The shortlist Products filed under the label are compared together Example Co.'s buyersOther point-of-care testing products
  2. Compare The tests The label sets the tests buyers judge on Example Co.'s buyersHow long results take, which tests it runs, cost per test, staff time
  3. Pay The budget Money comes from a line that already exists Example Co.'s buyersThe same budget line as their other lab supplies

No familiar label: a product loses all three jobs at once. Buyers struggle to find it, compare it and fund it.

Example Co.'s 12 calls and the words heard in them are fictional. The three jobs a label does (find, compare, pay) are my framing, not research.

Do this

Collect the exact words from your last ten sales calls and your search data, and write the top three labels on one line.

Step 2: Decide whether the existing label helps or hurts you

An existing category is free distribution. Buyers already know what it is for, directories and analysts already list it, and budgets already exist. It hurts you only when its tests hide your real advantage, for example when a product that changes how people work is judged on price alone.

A new label must also pass a test of its own. Jade Lo, Peer Fiss, Eunice Rhee and Mark Kennedy argue that viable categories are those "found useful for sensemaking, analysis, and coordination because they balance both coherence and distinctiveness"[7]. In plain words, the members of a category must be similar enough to each other, and different enough from other categories, to help people make sense of the market. If no other product could ever sit under your label, buyers have nothing to compare. If it looks like an old label, it adds nothing.

Figure 2 A viable category is alike inside and distinct outside

Hard to tell apart from other categoriesLow distinctiveness

Clearly different from other categoriesHigh distinctiveness

Members alikeHigh coherence

Members alike, but hard to tell apart from other categories Fails one test Looks like an old label It adds nothing: buyers already have a word for it Members alike, and clearly different from other categories Passes both Zone of viability Useful for sensemaking, analysis and coordination

Members share littleLow coherence

Members share little, and hard to tell apart from other categories Fails both Neither alike nor distinct The label gives buyers no reason to use it Members share little, but clearly different from other categories Fails one test Nothing to compare The extreme: a category of one, where no other product could ever fit

Lo, Fiss, Rhee and Kennedy (2020), Academy of Management Review: a theory paper, not a measurement. They call the balanced area a zone of viability. The wording in the four cells is my plain-words reading.

In medical products, a label is also a claim. The FDA's rule on "intended uses" says the objective intent of the people responsible for a product's labeling "may, for example, be shown by labeling claims, advertising matter, or oral or written statements by such persons or their representatives"[8]. So a label that suggests more than the product is cleared to do is a regulatory question as well as a marketing one.

If you sell an AI product. In my view, "AI" in a category label is a claim buyers will test, not decoration. Expect them to try the product on their own data and judge it on accuracy and on what happens to that data. A label like "AI lawyer" or "autonomous" sets the bar your product must clear in that trial. Regulators can read it as a claim too. In 2025 the FTC finalized an order against DoNotPay, which had "promoted its online subscription service as 'the world's first robot lawyer'". The FTC alleged that "the company did not test whether its 'AI lawyer' operated to the level of a human lawyer", and the order "prohibits DoNotPay from advertising that its service performs like a real lawyer unless it has sufficient evidence to back it up"[9]. That is what the FTC alleged and ordered, not a finding about any other company. AI labels also age fast: a word that sounds new this year can sound generic next year. So my advice is to name the category by the job the buyer pays for, and use "AI" to explain how you do it.

Three questions, asked in order, settle most cases.

Figure 3 Create a new category only when the old label works against you
  1. 1 Under the label buyers use now, are you judged on tests that hide your advantage? Example Co.: partly. Its in-visit fit is hidden Yes If no Win the existing category with a sharp point of difference
  2. 2 Does your product change how buyers work or budget, not just do it better? Example Co.: yes. Routine lab tests fit into an ordinary visit Yes If no Create a subcategory: the old label plus your difference
  3. 3 Can you fund years of teaching the market before rivals copy the label? Example Co.: not yet Yes If no Choose the subcategory now and revisit a new label later. Example Co. stops here.

All three yes: create the new category, and budget for the teaching it needs.

Answer in order; a no stops you at that question's If no box. My judgement, not research. Example Co.'s answers are fictional.

Example Co., step 2 Fictional

Under "point-of-care testing", Example Co. is compared on turnaround, test menu and cost per test, tests it can win. What the label hides is the part practices value most: routine lab tests that fit into an ordinary primary-care visit, run by a nurse or medical assistant. The team answers "partly" to the first question, "yes" to the second and "not yet" to the third.

Do this

Answer the three questions in writing, with one piece of evidence for each answer. If your product is regulated, check the label against what it is cleared for.

Step 3: Check who the new label is for

A new label lands differently with different audiences. Elizabeth Pontikes studied US software organizations from 1990 to 2002. Organizations that claimed ambiguous labels were "less appealing to consumers, an audience of market-takers, but more appealing to venture capitalists, who are market-makers"[10]. Market-takers use labels to find and judge products. Market-makers want to redraw the market. Even investors differed: independent venture capitalists preferred ambiguous labels, while corporate venture capitalists avoided them.

A later study of the same industry found that firms were more likely to introduce new category labels when their products combined existing categories, or when their category's meaning was ambiguous. It also found "an association between firms that use category innovation and their likelihood to IPO", which the author describes as preliminary evidence[11].

Figure 4 A new, ambiguous label put customers off but appealed to venture capitalists

A new, ambiguous category label

Customers Market-takers

Use labels to find and judge products.

Less appealing

Example Co. tells them, in sales calls and sales materials"point-of-care testing"

Venture capitalists Market-makers

Want to redraw the market structure.

More appealing

Independent VCs Acted as market-makers and preferred ambiguous labels

Corporate VCs Acted as market-takers and avoided them

Example Co. tells them, in its investor deck"in-visit diagnostics"

Pontikes (2012), Administrative Science Quarterly: a longitudinal study of US software organizations, 1990 to 2002. Example Co.'s split by audience is fictional.

In my view, the practical lesson is simple. The story that excites a seed investor can confuse the buyer you need this quarter. You can tell both stories, as long as each audience hears the one that helps it decide.

Example Co., step 3 Fictional

Practice managers and physician-owners are market-takers: they want a tool they can file, compare and buy this quarter. Investors are market-makers. Example Co. leads with "point-of-care testing" in sales calls and on its sales materials, and keeps the bigger "in-visit diagnostics" story for its investor deck.

Do this

List your three most important audiences, mark each as a market-taker or a market-maker, and match the label to each.

Step 4: If you create a category, budget for two jobs

Creating a category is two jobs, in order. First you make the category itself credible. Then, once buyers accept it, competitors arrive and you must stand out inside it. Chad Navis and Mary Ann Glynn tracked this in US satellite radio from 1990 to 2005: "the legitimation of a new market category precipitates shifts in the focus of market actors' attention from the category as a whole to the differentiation of firms within"[3]. The teaching you pay for in the first job also helps the rivals who arrive in the second.

Figure 5 Once satellite radio was accepted, attention moved from the category to the firms inside it
Before the licenses Emergence 2002 Early growth 1990 1997 2002 2005
  • Before the licenses 1990 to 1996
  • Emergence 1997 to 2001
  • Commercialization 2002
  • Early growth 2003 to 2005

1990: Satellite CD Radio petitions the FCC for spectrum. 1997: the FCC auctions licenses to two firms, later XM and Sirius.

Emergence, 1997 to 2001

Job 1: make the category credible

Firms "promoted a collective identity, muting their individual distinctiveness" Analysts "focused on the collective identity of the satellite radio category"

2002Legitimacy threshold

Early growth, 2003 to 2005

Job 2: stand out inside it

Firms "a new willingness by the firms to claim distinctiveness from each other" Analysts "placed greater emphasis on the individual firms as separate and distinct actors" Show the numbers as a table
Period Years Where attention went
Before the licenses 1990 to 1996 A 1990 petition to the FCC; licenses auctioned in 1997
Emergence 1997 to 2001 The category as a whole
Commercialization 2002 The legitimacy threshold
Early growth 2003 to 2005 Differences between the two firms

Navis and Glynn (2010), Administrative Science Quarterly: US satellite radio over its first 16 years. The periods and the quoted phrases are the paper's; the job names are mine. Drawn to scale in years.

What Salesforce's and HubSpot's filings show

Two well-known software companies introduced new labels, and their IPO filings show how they framed them. These are what the filings say, not an account of how either company made its choices internally.

Salesforce: "on-demand", sold into CRM. Its 2004 prospectus describes "application services that allow organizations to easily share customer information on demand" and explains that "On-demand application services enable businesses to subscribe to a wide variety of application services that are developed specifically for, and delivered over, the Internet on an as-needed basis". The new label was tied to a budget buyers already had: "We believe that the CRM applications market, which was approximately $7.1 billion in 2002 according to a July 2003 report by IDC, is one of the first to benefit from on-demand application services." The prospectus also disclosed the risk: "The market for on-demand application services is new and unproven"[1].

HubSpot: "inbound", sold as marketing and sales software. Its 2014 S-1 opens: "We provide a cloud-based marketing and sales software platform that enables businesses to deliver an inbound experience." It defines the idea against the old way: "more relevant, more helpful, more personalized and less interruptive than traditional marketing and sales tactics." It taught the idea at scale: "Our founders, Brian Halligan and Dharmesh Shah, wrote the best-selling marketing book Inbound Marketing: Get Found Using Google, Social Media and Blogs", and its blog "attracts approximately 1.5 million visits each month". It named the risks too: "If we fail to maintain our inbound marketing thought leadership position, our business may suffer", and "The market for inbound marketing products is still evolving"[2].

Figure 6 Both new labels were sold into budgets buyers already had

Salesforce Prospectus, June 2004

New label"on-demand application services"

Sold into "the CRM applications market" Promise Reduce "many of the traditional expenses and complexities of enterprise software implementations" Risk it disclosed The market is "new and unproven"

$7.1 billion the CRM applications market in 2002, according to a July 2003 IDC report the prospectus cites

HubSpot Form S-1, August 2014

New label"an inbound experience"

Sold as "a cloud-based marketing and sales software platform" Promise "less interruptive than traditional marketing and sales tactics" Risk it disclosed The market is "still evolving"

1.5 million visits a month to HubSpot's blog, part of how it taught the idea

Show the numbers as a table
Filing Number What it measures
Salesforce prospectus, 2004 About $7.1 billion The CRM applications market in 2002, per IDC
HubSpot S-1, 2014 About 1.5 million Visits to HubSpot's blog each month

Quoted from Salesforce's prospectus of 22 June 2004 and HubSpot's Form S-1 filed on 25 August 2014. What the filings say, not how either company decided internally.

The pattern I take from both: the new word described a better way to buy something buyers already paid for. Neither filing asks buyers to find money for an unknown thing.

An AI example: what C3.ai's filing shows

C3.ai: "Enterprise AI", set against legacy software. The prospectus summary in C3.ai's 2020 S-1 opens: "C3.ai is an Enterprise AI software company." It states the aim: "to establish and maintain a global leadership position in Enterprise AI." And it names the teaching job: "Market acceptance of an Enterprise AI solution depends in part on market awareness of the benefits that Enterprise AI can provide over legacy products, emerging point products, and manual processes"[12]. This is what the filing says, not how the company decided internally. In my view it fits the same pattern: the new label is defined against software buyers already pay for, and the filing ties acceptance to teaching buyers what the label means.

Example Co., step 4 Fictional

A true new category would mean persuading practices, then health systems, that "in-visit diagnostics" is something they need, before they ever compare vendors. In a health system, a value analysis committee (the group that reviews new products before the system buys them) would also ask for evidence. With a seed-stage team of nine and a two-quarter target of 25 clinics signed (fictional), Example Co. cannot fund both jobs yet.

Do this

If you still want a new category, write down what you will spend each quarter on teaching the market, and what you will do when the first competitor adopts your label.

Step 5: Choose your path and write the category line

Most startups land on one of three options. Enter the existing category and win on a sharp difference. Create a subcategory: the old label plus a new qualifier, which keeps the budget and adds a reason to choose you. Or create a new category, with a new label and the teaching budget it needs. Write the choice as one line your whole team uses: "For [buyer], we are the [category] that [difference]."

Figure 7 Example Co. keeps a label buyers know and adds its difference
Easier to find and compare More distinct, more to teach

Easier to find and compare

  1. Existing category "Point-of-care testing" Compared on old tests; the in-visit fit is hidden
  2. Subcategory "Point-of-care testing for the primary-care visit" Old budget, new reason to choose Example Co.'s choice
  3. New category "In-visit diagnostics" Distinct, but buyers must learn what it is first

More distinct, more to teach

Example Co.'s category line

  • independent primary-care practices
  • point-of-care analyzer
  • gives routine blood test results during the visit, not two days later
  • Buyer who it is for
  • Category the shelf buyers know
  • Difference the reason to choose it

The template: "For [buyer], we are the [category] that [difference]."

Three ways to describe the same product. Example Co.'s choice and its category line are fictional; check any label against the product's cleared use.

Example Co., step 5 Fictional

The team chooses the subcategory. Its category line: "For independent primary-care practices, Example Co. is the point-of-care analyzer that gives routine blood test results during the visit, not two days later." "In-visit diagnostics" stays as a description of the product, not its category, and the team checks the line against the product's cleared use before it goes on the website. It will revisit the choice when it starts selling to health systems, where buyers judge products on different tests, such as published accuracy evidence and a link to the electronic health record.

Do this

Write your category line, then test it on five buyers. Can each of them say what you are and what you replace, in one sentence?

Free tool

Draft your category line

The free Positioning Statement Generator turns seven guided prompts into a positioning statement and flags vague words, so you can draft and refine the line from Step 5.

Draft the line

What changes by stage?

  • Pre-seed — Use your buyers' existing words. You are still learning which category you are really in.
  • Seed — Choose between the existing category and a subcategory. Test the category line in sales calls and on your homepage, and keep any new label as a qualifier.
  • Series A–B — If your product clearly combines categories, or your category is poorly defined, a new label may now pay off. Budget for the teaching, and plan for the moment buyers start comparing firms inside your category.

What mistakes should I avoid?

  • Inventing a label to sound bigger. A label nobody searches for is a label nobody finds.
  • Telling every audience the same story. Buyers and investors react to new labels differently.
  • Forgetting the second job. Once a category is accepted, attention moves to the firms inside it.
  • A category of one. If no other product could ever fit, buyers have nothing to compare.
  • A label that claims more than the product is cleared for. In medical products, marketing words can be read as intended use.
  • Never revisiting the choice. Your category can change when you move to a new segment.

Related on AAJ

Articles

Playbooks

Templates and tools

References

The superscript numbers in the article match this list.

  1. salesforce.com, inc. (2004). Prospectus (Form 424B1), 22 June 2004. US Securities and Exchange Commission. sec.gov
  2. HubSpot, Inc. (2014). Registration statement on Form S-1, filed 25 August 2014. US Securities and Exchange Commission. sec.gov
  3. Navis, C. and Glynn, M. A. (2010). "How New Market Categories Emerge: Temporal Dynamics of Legitimacy, Identity, and Entrepreneurship in Satellite Radio, 1990–2005." Administrative Science Quarterly, 55(3), 439–471. doi:10.2189/asqu.2010.55.3.439
  4. Zuckerman, E. W. (1999). "The Categorical Imperative: Securities Analysts and the Illegitimacy Discount." American Journal of Sociology, 104(5), 1398–1438. doi:10.1086/210178
  5. 42 U.S.C. § 263a(d)(3), Certification of laboratories: the examinations and procedures that may be waived. Legal Information Institute, Cornell Law School. law.cornell.edu
  6. US Food and Drug Administration. "CLIA Waiver by Application." In Vitro Diagnostics regulatory assistance. fda.gov
  7. Lo, J., Fiss, P. C., Rhee, E. and Kennedy, M. T. (2020). "Category Viability: Balanced Levels of Coherence and Distinctiveness." Academy of Management Review, 45(1), 85–108. doi:10.5465/amr.2017.0011
  8. US Food and Drug Administration. 21 CFR 801.4, "Meaning of intended uses". ecfr.gov
  9. US Federal Trade Commission (2025). "FTC Finalizes Order with DoNotPay That Prohibits Deceptive 'AI Lawyer' Claims, Imposes Monetary Relief, and Requires Notice to Past Subscribers." Press release, 11 February 2025. ftc.gov
  10. Pontikes, E. G. (2012). "Two Sides of the Same Coin: How Ambiguous Classification Affects Multiple Audiences' Evaluations." Administrative Science Quarterly, 57(1), 81–118. doi:10.1177/0001839212446689
  11. Pontikes, E. G. (2022). "Category innovation in the software industry: 1990–2002." Strategic Management Journal, 43(9), 1697–1727. doi:10.1002/smj.3383
  12. C3.ai, Inc. (2020). Registration statement on Form S-1, filed 13 November 2020. US Securities and Exchange Commission. sec.gov

More in Strategy & Positioning

Part of the Strategy & Positioning hub - see all 19 resources on this topic.

Also useful in Brand & Voice.