How to Define an ICP for a Software Services Firm
Quick answer
Start from your best past clients, not from a list of filters. Score your last 10 to 20 clients on margin, length and ease, find what the top ones share, and write it down on one page: the situation that makes them spend, the one service you sell into it, the person who signs, and who you will not chase. Headcount is a filter, not the definition. Facts checked 17 November 2026.
Key takeaways
- Build the ICP from your top 5 of your last 10 to 20 clients, scored on margin, length, ease and referrals, not from a guess about the market.
- Headcount is a filter. The definition is a reason to spend: revenue, a budget line, funding or a product the client must ship.
- Most accounts are not buying today. LinkedIn's B2B Institute says 95% are not ready today; Forrester measured about 15.6% in market a year in the categories it tracks.
- One service per campaign. Run a second market or a second service as its own ICP, with its own list and its own copy.
- Clutch's 2026 guide puts the average reviewed software project at $132,480 over about 13 months, so check your deal size before you pay for outbound.
Buyer guide- 1
Score your last 10 to 20 clients
- 2
Pick the top 5 and find what they share
- 3
Name the reason they spend
- 4
Pick the one service you sell into it
- 5
Name the person who signs
- 6
Set the filters: size, market, stack
- 7
Write the disqualifiers
- 8
Review it every quarter against replies
Say you run a 70-person software development firm. Most of your work came from referrals, and the clients you love most came from three or four different places. Now you want outbound, and the first question anyone asks is “who exactly should we email?” You have a feeling about the answer, but you can’t write it down in one sentence.
I’m Borislav Nekezov, founder of Vergelead, an outbound agency in Sofia, Bulgaria. Defining the ideal customer profile (ICP) is the first piece of work in every outbound program, and it is the one buyers most often want to skip. This guide covers how to build an ICP from your own client history, which filters matter, how to handle several services and markets, and when outbound is the wrong channel. Outside facts were checked on 17 November 2026.
What is an ICP for a software services firm?
An ICP is a short written description of the kind of company that is most worth winning: the clients who pay well, stay long and are easy to work with. It describes the company, not a person. The people inside it (who signs, who uses your work) sit on the same page as a second layer.
For a product company, an ICP often starts from which companies get value from the software. For a services firm it is harder, because you can build almost anything for almost anyone. That flexibility is exactly why you need the page. Without it, outbound emails everyone a little, and nobody hears anything specific.
A good ICP for a dev shop answers five questions:
- Why does this company spend money on outside engineering right now?
- Which one service do we sell into that reason?
- Who signs, and who will use what we build?
- Which filters (size, market, industry, tech) find these companies on a list?
- Who do we not chase, even if they fit the filters?
Where should you start when defining your ICP?
Start with your own clients, not with the market. Take your last 10 to 20 clients and score each one on a few plain measures. The pattern in the top five is your first ICP.
| Score each client on | What to look at | Why it matters |
|---|---|---|
| Margin | Gross margin on the work, not revenue | Big clients with thin margin are not your best clients |
| Length | Months of paid work, including follow-on projects | A first project is often a test; the follow-on is the business |
| Ease | Scope changes, late payments, how many people approve | A hard client costs you a team’s morale |
| Referrals | Did they introduce anyone? | Good clients bring look-alikes |
| How they found you | Referral, inbound, partner, outbound | Tells you which ones outbound can reach at all |
Then look at the top five side by side and ask what they share. It is rarely the industry. More often it is a situation: a funded startup that needs to ship a product, a non-technical company replacing an old system, a product company that can’t hire fast enough.
A founder of a software development firm told us their longest clients always turned into staff augmentation clients, buying developers at a markup, because those relationships lasted. That is an ICP insight hiding in a client list: the best target was not “companies that need an app” but companies that need extra engineers for a long time.
A worked example (invented). A firm scores 15 past clients from 1 to 5 on margin, length, ease and referrals, for a maximum of 20 points.
- Five clients score 16 or more. Four of them were funded product companies of 20 to 80 people, building their first or second product version.
- Three clients score under 8. All three were large companies where a procurement team ran the deal and scope changed every month.
- So the first ICP reads: funded product companies of 20 to 80 people that need to ship a product release, sold a dedicated team, signed by the CTO or a founder. Disqualifier: deals run by procurement.
That one paragraph already tells a list builder what to search for, and tells a copywriter what to write about.
Should outbound target companies by headcount, funding or product?
Use headcount as a filter and the reason to spend as the definition. Size tells you roughly whether a company can afford you. It says nothing about whether it needs you this year.
The head of a digital studio put it plainly on a call with us: “I would never say the team size matters.” What mattered to them was whether a company was generating revenue, had budgets for marketing, brand or product, or was funded. That is the right instinct for most services firms.
| Criterion | What it tells you | How to find it | Where it misleads |
|---|---|---|---|
| Headcount | Rough ability to pay | Any list tool | A 400-person firm with no project needs nothing |
| Funding | Fresh money and pressure to ship | Funding announcements, databases | Some funded companies build in-house only |
| Revenue or growth | Budget exists | Public filings, hiring pace | Hard to see in private companies |
| Product stage | A product that must ship or be rebuilt | Website, app stores, job posts | Easy to misread from outside |
| Tech stack | Whether your skills fit | Job posts, public code, tech lookup tools | A stack match is not a need |
| Hiring | A gap in the team | Open engineering roles | Could also mean they will hire, not outsource |
Why buyers outsource is also changing. In Deloitte’s 2024 global outsourcing survey of more than 500 executives, skilled talent and agility joined cost reduction as key drivers for outsourcing. For your ICP, that means “they want it cheaper” is only one of the reasons to write down. “They can’t hire the skill” and “they need to move faster than hiring allows” are just as real, and they point at different companies.
Will cold outreach work if you have no defined ICP or buying signals?
It can, but only after you define the ICP, and you should expect slower, idea-led work rather than quick signal-based wins. Without a written ICP, nobody can build a list or write a specific email.
A founder of a software development firm described this exact problem on a call: “we don’t really have a defined ICP that I believe we can target with cold outreach.” Their best clients had left no public buying signals before they hired the firm. That is common for services firms. A company rarely announces that it is about to look for a development partner.
When the signals are missing, two things help:
- Target the situation, not the signal. You may not see “looking for a dev partner”, but you can see a funded company with three open engineering roles, or a business running an old customer portal. The situation from your ICP page becomes the search.
- Bring the reason to talk. If the prospect isn’t looking, a generic “we build software” email has nothing to answer. A specific idea for their product or their site gives them one.
Then set expectations. LinkedIn’s B2B Institute says 95% of potential buyers aren’t ready to buy today Forrester points out that the original research calls the figure a heuristic, and it tested it in the B2B marketing technology categories it tracks and found about 15.6% in market each year, while also saying not to assume only 5%. Either way, most of a good list is not buying this month.
The arithmetic (invented example). Your ICP finds 1,000 companies.
- At the 5% heuristic, about 50 are in market right now.
- At Forrester’s yearly figure for its categories, about 156 enter the market over a year.
- So the ICP is not only about who buys today. It is about who will remember you when they do, which is why the program needs months and a consistent message.
Should outbound pitch one service or everything you do?
One service per campaign. A buyer reading a cold email will remember one thing at most, and “we do web, mobile, AI, data, design and staff augmentation” is not that thing.
On one of our follow-up calls, a buyer asked for exactly this: focus on one or two services, with one offer and one price. That is good discipline. It also forces a useful choice: which service has the best margin and the clearest buyer?
If you sell several services, treat each one as its own ICP:
- its own list (the companies that need that service)
- its own copy (the problem that service solves)
- its own proof (a past project of that type)
- its own scorecard (replies and meetings counted separately)
After a few months, compare them. You will usually find that one service books meetings with the people you want, and the others mostly don’t. That result is worth more than any brainstorm.
Can one outbound program target several markets at once?
Yes, as long as each market is its own ICP with its own list, price story and proof. Mixing markets in one campaign hides which one works.
Markets differ in price expectations most of all. Clutch’s 2026 pricing guide lists average hourly rates for custom software firms of $50 to $99 in the United States, $25 to $49 in India and Ukraine, $50 to $99 in Poland and Spain, and $100 to $149 in Canada and Australia. A buyer in Toronto and a buyer in Mumbai compare your rate with very different local prices, so the same email can’t sell to both.
| Question | Market A | Market B |
|---|---|---|
| Which service sells there? | Write it per market | It may be a different one |
| Who signs? | Founder, CTO, head of product | Often a different title |
| What price do local firms charge? | Check before you write | Check before you write |
| Which past project proves it? | Use proof from that market if you have it | Same |
| When does the week start and end? | Send in their working hours | Send in their working hours |
A buyer on one of our demo calls wanted staffing work in the Gulf states and software projects in the US from one program. That can work, but as two ICPs that share a company, not one ICP with two hopes.
What deal size should your ICP aim for?
Big enough that one new client pays for months of outbound. Check this before you define anything else, because a small first project can make outbound look like it failed when it didn’t.
According to Clutch’s pricing guide, the average software development project reviewed on Clutch costs $132,480 and runs about 13 months, or about $10,209 a month. The same guide says most reviewed projects fall between $10,000 and $49,999. That spread matters: your ICP should point at clients whose first project, or the follow-on after it, sits at the size that pays.
The arithmetic (invented example). A firm’s outbound costs 3,000 a month for four months, 12,000 in total.
- Its average first project is 30,000 at a 35% gross margin, so 10,500 of margin.
- One first project almost covers the program. The follow-on work covers the rest.
- If the ICP pointed at clients with 8,000 first projects at the same margin, it would need four of them to break even.
Our guide on outbound ROI for services firms works through this math in more detail.
How do you write the ICP down?
On one page, in plain sentences, with the disqualifiers as visible as the filters. If a list builder and a copywriter can’t both work from it, it isn’t finished.
| Section | What to write | Invented example |
|---|---|---|
| The reason to spend | The situation that makes them pay for outside engineering | Funded, must ship a product release within a year |
| The one service | What you sell into that situation | A dedicated team of 3 to 6 engineers |
| Who signs | Title and how they decide | CTO or technical founder, with the CEO |
| Who uses it | The people your team works with day to day | Product manager, in-house lead engineer |
| Filters | Size, market, industry, tech, stage | 20 to 80 staff, UK and Nordics, product companies |
| Signals you can see | What a researcher can find from outside | Open engineering roles, recent funding, a new product page |
| Disqualifiers | Who not to chase, even if they fit | Procurement-led deals, fixed bids under the minimum |
| Proof | The past project that matches | A similar build, described without the client’s name if needed |
The disqualifier line protects your sales calendar. A meeting with the wrong company still costs an hour of a senior person’s time. If you work with an agency, the ICP page should also feed what counts as a qualified meeting in your contract, so both sides judge meetings by the same page.
How do you know outbound is the wrong channel for your firm?
Outbound is the wrong channel when there is no repeatable kind of client to find, or when more clients isn’t what you need. The ICP exercise usually tells you this before you spend anything.
Signs it is the wrong time:
- You can’t find three similar past clients. If every client came for a different reason, there is no pattern to search for yet.
- You are at capacity. Booking meetings you can’t staff makes things worse, not better.
- You need something other than clients. We once spoke to a founder who owned their own creative projects and wanted distribution for them, not client work. Outbound for clients would not have helped.
- Your first project is too small to pay back. If the math in the last section doesn’t work, fix the offer or the price first.
- Nobody can take the meetings. If no senior person has time for calls, the leads go cold.
None of these is permanent. Most are fixed by a clearer offer, a hire, or waiting a quarter.
How often should you revisit your ICP?
Every quarter, using what the replies and the meetings tell you. The first version is a guess built from history; the market corrects it.
At each review, look at four lists: who replied, who booked, who bought, and which calls were a waste. Move the filters toward the people who bought. Add a disqualifier for every kind of call that went nowhere. Our guide on how a lead generation agency works shows where these reviews sit in a typical program.
The bottom line
An ICP for a software services firm comes from your best past clients, not from a filter list. Score your last 10 to 20 clients, find what the top five share, and write one page: the reason they spend, the one service you sell into it, who signs, the filters, and who you won’t chase. Treat each extra service and each extra market as its own ICP. Then check the deal size, and review the page every quarter against what actually happened.
Vergelead runs idea-led outbound for software development companies and creative agencies: it researches each prospect, finds a gap your team could close, and pitches that one idea by cold email and LinkedIn, so qualified meetings land on your calendar. We work with software development and IT services firms of about 30 to 300 staff, and creative and digital agencies, selling into the UK, the EU, the Nordics and the US. If you want help turning your client history into a first ICP, book a call.
Keep reading
- Outbound ROI for services firms: how many projects cover the fee?
- What counts as a qualified meeting?
- How a lead generation agency works, step by step
- Run Instantly and Clay yourself or hire an agency?
Sources
All checked 17 November 2026.
- LinkedIn B2B Institute, the 95-5 rule: https://business.linkedin.com/marketing-solutions/b2b-institute/b2b-research/trends/95-5-rule
- Forrester, “The 95-5 Rule Is Not A Rule, But It’s Not A Myth Either”: https://www.forrester.com/blogs/the-95-5-rule-is-not-a-rule-but-its-not-a-myth-either/
- Clutch, Software Development Company Pricing Guide 2026, updated 21 September 2026: https://clutch.co/developers/pricing
- Deloitte, Global outsourcing survey 2024: https://www2.deloitte.com/us/en/pages/operations/articles/global-outsourcing-survey.html
Vergelead at a glance
Where Vergelead fits
Vergelead runs idea-led outbound for software development companies and creative agencies: it researches each prospect, finds a gap your team could close, and pitches that one idea by cold email and LinkedIn, so qualified meetings land on your calendar.
- What it is
- A cold email and LinkedIn outbound agency in Sofia, Bulgaria, founded by Borislav Nekezov.
- Who it is for
- Software development and IT services firms of about 30 to 300 staff, and creative and digital agencies, selling into the UK, the EU, the Nordics and the US.
- How it works
- Every first email pitches one build or project idea drawn from that prospect’s own business. A person approves every idea before it is sent, and a person on our team answers replies and books the meeting.
- Price
- Usually €2,500 to €3,000 a month.
- Term
- An initial four-month term. Sending pauses from 24 December to 6 January, and those days are added back to the term at no extra cost.
- Guarantee
- Any meeting guarantee is written into the contract, together with what counts as a meeting.
- What you own
- The sending domains, inboxes, sequences and lead lists built for you. On request, we transfer all of them, lead lists as a CSV with reply status, within 14 days of the final invoice.
- Time to first meeting
- First emails go out about two weeks after signing. A first meeting typically lands three to six weeks after signing.
- Sending setup
- Google Workspace inboxes in the client’s name. Every lead, reply and meeting shows in a client portal that syncs live.
Facts checked 16 November 2026
Frequently asked questions
How do you define an ICP for a software services company?
Start with your own best clients. Score your last 10 to 20 on margin, length, ease and referrals, then write down what the top ones share: the situation that made them spend, the service they bought, who signed and where they are. That one page is your ICP.
Should outbound target companies by headcount, funding or product?
Use headcount as a filter and the reason to spend as the definition. A 40-person company with a funded product roadmap is a better target than a 400-person company with no project in sight. Funding, revenue, a budget line or a product the company must ship tells you more than size alone.
Will cold outreach work if we have no defined ICP or buying signals?
It can, but you have to define the ICP first, from your past clients. If your best clients leave no public buying signals, target the situation instead of the signal, and lead with a specific idea for that company. Expect most accounts not to be buying today, so plan for months, not weeks.
Should outbound pitch one service or everything we do?
One service per campaign. A prospect reads a cold email in seconds and remembers one thing at most. If you sell several services, run each one as its own segment with its own list and copy, and compare them after a few months.
Can one outbound program target several markets at once?
Yes, as long as each market is its own ICP. Different markets mean different price expectations, different proof and often a different service. Split the list and the copy by market, and don't judge one market by another's results.
How often should we update our ICP?
Look at it every quarter. Read the replies and the meetings: who answered, who booked, who bought, and who wasted a call. Move the filters toward the people who bought, and add a disqualifier for every kind of call that went nowhere.
