Mazo › Go-to-market strategy for B2B SaaS › ICP for horizontal SaaS

Guide · Horizontal B2B SaaS

How to choose an ICP when your SaaS could serve anyone

By Madalena Rugeroni, founder of Mazo · Updated 15 September 2026 · 8-minute read

Short answer: if your SaaS could serve anyone, pick one segment anyway. Score candidates on urgency, ability to pay, reachability, word of mouth and your own access. Check the winner against who already activates and stays, then aim your homepage, onboarding and outreach at that segment only. Expand once the first segment sells repeatably.

Why "everyone" means no one converts

A horizontal product — a form builder, a client portal, a scheduling tool — can technically serve any business. Its go-to-market can't. Every part of it needs a specific buyer to aim at:

Positioning. April Dunford's process starts from the alternatives a customer would use without you. A law firm and a marketing agency replace different things, so "for everyone" can't say what you replace.
The homepage. Des Traynor's rule: one product, one primary job. A homepage that lists six jobs gets hired for none.
Channels. Outbound needs companies you can name, search needs a term people type, community needs a room your buyers are in. "Everyone" has none of these.
Pricing. Madhavan Ramanujam's point is that two similar-looking companies can value the same product 10x apart. One price for everyone undercharges the buyers who value you most.

It's the first of the six recurring GTM problems Mazo diagnoses: an ICP that exists on paper but isn't enforced, so demos aren't tailored and the website speaks to everyone.

Choosing a segment doesn't mean rebuilding the product. Christopher Lochhead's practical middle path is to niche down an existing category — "CRM for construction" — rather than invent a new one.

Mazo's rule of thumb: keep the product horizontal and make the go-to-market vertical. One segment on the homepage, in onboarding and in outreach, until that segment sells repeatably.

How to pick your beachhead segment

List three to five candidate segments, then score each from 1 to 5 on five criteria.

Criterion	Question to ask	Evidence that counts
Urgency	Is there a moment that forces them to act now?	Trigger events you can observe (Winning by Design's in-market signals), plus a workaround or money already spent (Cindy Alvarez)
Ability to pay	Do they have budget, and do they value the outcome highly?	Existing spend on the problem; willingness-to-pay answers compared by segment (Ramanujam)
Reachability	Can you find them cheaply?	You can list 50 dream accounts by hand, or they gather in one community, event or platform
Word of mouth	Do they talk to each other about tools?	Peer communities; product use that shows you to non-users (Brian Balfour's growth loops); "How did you hear about us?" answers (Chris Walker)
Your unfair access	Can you get meetings this month?	Your network, industry background and credibility. Lenny Rachitsky notes most successful B2B companies got their first 10 customers this way

Add up the scores, with one veto: a segment that scores low on urgency is out, whatever its total. Alvarez's rule is that no workaround means no felt pain, and no felt pain means nobody buys.

Then check Balfour's model–market fit: are there enough buyers in this segment, at your price, to reach your next revenue goal? If you need 400 customers and the segment holds 300 companies, pick a bigger segment or charge more.

Mazo's rule of thumb: pick the segment with the highest urgency that you can reach this month, not the biggest market. A small segment that buys fast gives you customers, case studies and referrals. A big one that doesn't just gives you traffic.

Use your sign-up data to find the segment

If you already have users, don't guess. Dunford's best-fit customers are the ones who bought fastest, stayed longest and got the most value; whatever they share is your segment. Casey Winters adds the retention view: when a cohort curve flattens, product-market fit exists for the users who stayed, so find out who they are and aim your ICP and acquisition at them.

Tag every sign-up by segment. Add one required sign-up question (company type or use case) and back-fill existing accounts from email domains and onboarding answers.
Compare segments on outcomes, not sign-ups: activation, trial-to-paid, still paying after three months, and expansion.
Read churn by segment. Lincoln Murphy's bad-fit signs are customers who churn inside 90 days and the same profile churning again and again.
Read "How did you hear about us?" by segment. A segment that arrives through referrals already has word of mouth.
Interview five best-fit accounts about the moment they started looking. That trigger is your urgency signal.

With fewer than 10 customers there's no pattern to find yet. Treat your segment as a hypothesis and test it in five interviews with one profile.

Mazo's rule of thumb: don't pick the segment with the most sign-ups by default. Pick the one with the best activation and retention, even if it's smaller. As Elena Verna puts it, acquisition scales whatever retention you already have.

Positioning per segment without rebuilding the product

Positioning is context, not code. Run Dunford's sequence for your chosen segment only: their alternatives, what you do that those can't, the value that creates for them, and only then the category frame. The job and the thing it replaces change by segment (Traynor); the product mostly doesn't.

Change per segment

Homepage headline and the alternative you name
Onboarding templates and example data — Wes Bush's empty states that teach
Case studies and testimonials from that segment
Outreach list, trigger events and first line
Disqualifying questions for your anti-ICP (Winning by Design)

Keep the same

The core product and data model
One plan and one value metric, until pricing research says otherwise
The sign-up flow, apart from the segment question
Your brand

If you later run a second segment, give it its own landing page rather than one homepage that tries to hold both.

Mazo's rule of thumb: write one positioning sentence per segment — "We're a [category] for [segment] who need [outcome] without [the alternative's problem]" — and put only your beachhead's sentence on the homepage. If a sentence fits two segments equally well, it isn't specific enough.

When to expand to the next segment

Expand when the first segment is repeatable, not when it gets boring. The signs:

You've closed around 10 customers from the segment with the same pitch — Jason Lemkin's bar before handing sales to anyone else.
One channel brings that segment in, and you can measure it.
Retention in the segment has flattened, ideally with net revenue retention above 100%.
The segment can't reach your next revenue goal at your price: Balfour's model–market fit is running out.

Choose the next segment by adjacency: it should share the buyer and job, or the channel, with the one that works. Change one thing at a time. Balfour warns that changing one element of your model can silently break a fit two steps away.

And plan it before you need it. Andrew Chen's law of shitty clickthroughs says every channel decays as competitors find it, so the segment you depend on today will get more expensive to reach.

Mazo's rule of thumb: don't add a second segment until the first has around 10 customers won with the same pitch and one channel that works. Then pick the neighbour that shares your buyer or your channel, never one that needs a new buyer and a new channel at once.

Worked example (fictional)

PortalKit: a client portal with 1,000 sign-ups and no ICP

Fictional company, round numbers, for illustration only

PortalKit lets any service business share files, collect documents and send updates. Its sign-ups split like this:

Segment	Sign-ups	Activated	Paying	Paying after 3 months
Marketing agencies	400	120	20	10
Accounting firms	200	100	30	27
Coaches	300	60	10	4
IT consultants	100	40	5	4

Agencies look like the ICP because they sign up most. The data points to accounting firms: half activate, 15% pay, and 9 in 10 of those still pay three months later. Five interviews show why: every firm chases clients for documents before tax deadlines — a trigger event — using email and shared drives.

The decision: accounting firms score highest on urgency and ability to pay, and one co-founder used to work at a firm, which covers access. PortalKit rewrites its homepage as a client document portal for accounting firms, adds a "request documents" onboarding template and builds a list of 50 firms. Agencies stay the candidate next segment until the accounting pitch closes consistently through one channel.

Common mistakes when choosing an ICP for horizontal SaaS

Picking the segment with the most sign-ups instead of the one that activates and stays.
Choosing by market size. A big segment you can't reach is worth less than a small one you can.
Segmenting by demographics only. "10–200 employees" isn't a segment; the trigger that makes them buy is.
Rebuilding the product for a segment before the positioning has been tested.
Two segments on one homepage. Buyers from both leave unsure it's for them.
Expanding because growth feels slow, before the first segment is repeatable.

Tools for this

Write your segment down with the ICP worksheet, then test your segment-specific line with the positioning stress test. The willingness-to-pay test helps you compare what each segment will pay, and the first 10 customers plan turns your beachhead into a week-by-week list. For sequencing, use the 90-day go-to-market plan generator or the 90-day GTM plan template; the free GTM scorecard scores your ICP sharpness alongside the rest of your go-to-market. Terms like beachhead are in the GTM glossary.

Mazo is an AI go-to-market advisor for B2B SaaS founders from €0 to €1M ARR: it builds your plan from where you are today with proven SaaS playbooks, then runs it with you every week, for €99 a month.

Who wrote this

Madalena Rugeroni built Mazo. She's an ex-Googler, a startup advisor and investor, and runs a portfolio of internet companies. Before that, as Head of Growth at Amplemarket, she scaled a B2B SaaS to $10M ARR. LinkedIn

FAQ

What is a beachhead segment in SaaS?

The first narrow segment you aim your whole go-to-market at, from homepage to onboarding to outreach, before you expand. For a horizontal product it lets you win one group of buyers decisively instead of converting a few of everyone.

Does a horizontal SaaS need an ICP?

Yes. The product can stay horizontal, but positioning, channels and pricing all need a specific buyer. Christopher Lochhead's middle path is to niche down an existing category, like "CRM for construction", rather than market to everyone.

How do I find my ICP from existing users?

Tag sign-ups by segment and compare activation, trial-to-paid and three-month retention. April Dunford's best-fit customers, the ones who bought fastest and stayed longest, define the segment. With fewer than 10 customers, treat it as a hypothesis and test it in interviews.

When should a SaaS expand to a second segment?

When the first is repeatable: around 10 customers won with the same pitch, one channel you can measure, and flat retention. Then choose a neighbouring segment that shares your buyer or your channel, not one that needs both to change.

Build your go-to-market around one segment

Mazo builds your 90-day plan from where you are today, then runs it with you every week. €99 a month, 14 days free.

Start 14-day free trial Not ready? Score your go-to-market free, no account needed →

Sources and further reading. Operators and books referenced: April Dunford (Obviously Awesome), Des Traynor, Madhavan Ramanujam (Monetizing Innovation), Christopher Lochhead (Play Bigger), Winning by Design, Cindy Alvarez (Lean Customer Development), Brian Balfour, Chris Walker, Lenny Rachitsky, Casey Winters, Lincoln Murphy, Elena Verna, Wes Bush (Product-Led Growth), Jason Lemkin and Andrew Chen. These are published methodologies credited to their authors; Mazo is not affiliated with or endorsed by them.