How Software Development Companies Get New Clients

Software development companies often get their first clients from referrals, the founder's network and repeat work for existing accounts. That can carry a company for years. The limit is that nobody controls when the next referral arrives, or what kind of project it brings.

The rest of the menu is familiar: directory profiles such as Clutch or GoodFirms, partnerships with agencies or tech vendors, content and SEO, events, and outbound.

A dev shop that wants to keep winning custom projects needs at least one channel where it picks both the buyer and the timing. That is direct outreach to one narrow group of buyers on LinkedIn and by cold email, with the company's own case studies inside the messages.

Deal size is what makes this pay - when a single project is worth tens or hundreds of thousands of euros, a few signed contracts cover months of effort.

In May 2023, one campaign my team ran for a German software development client went like this: 9,000 contacted people -> 34 booked sales calls -> 4 signed contracts worth EUR 1,152,000 in total.

Where dev shops find clients, and which channels you control

Every channel a dev shop uses can be judged on two questions. Can you choose who the buyer is? Can you choose when the first conversation happens?

Directories deserve one extra note. On Clutch, paying sponsors appear higher on a directory page than companies that do not pay, although sponsorship does not change a company's place in Clutch's own ranking (checked September 2026).

A place near the top of a directory page can be bought, and the buyer still does the choosing.

None of this means dropping the other channels. It means knowing which one you can switch on when the calendar is empty, and building that one properly.

Why referrals stop being enough

Referrals are usually the warmest leads a dev shop gets. The buyer arrives half-sold, with trust borrowed from someone they already believe.

The weakness is the one in the table above - you cannot choose the industry, the company size or the timing of the next referral. A company that lives on referrals has a strong year when its clients happen to talk about it, and a thin one when they do not.

Referrals also hide a timing problem. In 6sense's 2025 Buyer Experience Report, a survey of more than 4,000 B2B buyers (vendor research), 94% of buying groups had ranked their preferred vendors before first contact with a seller.

They bought from that early favorite 77% of the time.

A referral puts you on that list. So does a relevant case study that reached the right person months before the project existed. Without either, the first you hear of a project may be a request for proposals where someone else is already the favorite.

Keep the referrals. The clients they brought you are also targeting data you already own, and the buyer section below starts from them.

The math when one project is worth $200,000

The number that decides how a dev shop should sell is its average project value. In my post on how I decide which market to target I used this example: "Now sell $200,000 IT projects instead, which is a normal average in the software and IT consulting work I see."

At that price, "One extra project a month is twelve a year and $2.4 million, assuming you can actually deliver twelve more projects". Delivery capacity is a real limit, so check it before you plan the outreach.

The other half of the sum is how many people you have to reach. From the same post: "it takes somewhere between 150 and 250 contacted people to produce one booked call, and where you land inside that range depends on the niche, the country and the channel."

Now add the step from call to contract. Two campaigns further down give a rough rate: one turned 29 booked calls into 5 contracts inside a single month, another turned 34 calls into 4.

That is about one signed project for every 6 to 9 booked calls. It is a planning figure from two campaigns, so treat it as a starting point for your own spreadsheet.

Say the goal is 12 new projects next year, one a month:

For a business selling $2,000 one-off jobs, contacting 27,000 people a year would be hard to justify. For a dev shop it is a different conversation. In the May 2023 campaign above, each person contacted was worth about EUR 128 in signed contracts, EUR 1,152,000 across 9,000 people, and more contracts were still in the pipeline.

Then count your market. If your chosen niche holds a few hundred companies with one plausible buyer each, a volume plan will run out of people within months. A small market needs more depth per account: more than one decision-maker at each company, more follow-ups, or a second country.

Who signs a custom software contract

Defining your ideal customer is the usual first step. For software work you can go further, because the buyer side has real titles.

As I put it in the same market post: "In IT and software work it is often a VP of IT strategy, a head of IT, a director of IT, sometimes a CTO, and they are not interchangeable across company sizes."

Do not pick the title in a lead tool because it sounds senior. Pull your past projects and write three things next to each one: who signed, how big the company was, and what they were trying to fix.

That short list gives you a map of which title buys at which company size. Then find lookalikes: companies with the same size, structure and problem as your paying clients. No competitor can copy that list, because nobody else has your closed deals.

Expect more than one person per account. A custom build is often signed off by several people, so reaching two titles at the same company is normal, as long as each person gets a message written for their role.

Pick one buyer and one offer

Web, mobile, AI, cloud, UX and QA, for startups and enterprises in every industry. I see that kind of menu every week, and outbound is where it hurts first.

The first cost is the list. Nobody can build a serious list of "companies that need software".

In my post on committing to one offer and one audience I wrote the alternative: "You can build a very good one for heads of engineering at software companies of 50 to 200 people in Germany, and the list is most of the campaign."

The second cost is proof. Three case studies from one narrow lane look like a track record. Ten case studies from ten industries look like a series of experiments.

Three made-up illustrations of what one line can look like for a dev shop:

• Custom dispatch and tracking software for freight forwarders in Germany, Austria and Switzerland, sold to the head of IT
• Extra engineering capacity for SaaS companies with 50 to 200 staff, sold to the head of engineering
• Rebuilds of aging in-house systems for mid-sized manufacturers, sold to the director of IT

Each one tells you who to put on the list, what the first message is about, and which of your projects to show.

Then use the projects. Put your past clients, results and case studies inside the outreach message itself. A buyer who sees a project like theirs in the second line has a reason to reply that no adjective can give them.

Four software development campaigns, with the numbers

These are four campaign results in software development and IT from my own records. The figures are as published at the time. Where a number was never published, the cell says so. Read each row on its own; the rows are not meant to be added together.

A few things stand out once they sit side by side.

The revenue sits in the deal size. Campaign 1's four contracts averaged EUR 288,000 each. Campaign 2's five averaged EUR 49,200. Both were outbound campaigns in software development, and the revenue per contract differed almost six times over, which is why the math section starts from project value.

The funnel is wide at the top and narrow at the bottom. In campaign 2: 392 replies -> 29 booked calls, about 7.4% of replies. That month ran on LinkedIn alone, and I walked through the follow-up sequence behind it separately.

Campaign 1 sat slightly above the planning range. Across both channels it took about 265 contacted people per booked call: about 238 on the cold email side and about 308 on the LinkedIn side, which had its own email follow-ups. That is a little above 150 to 250, and the month still paid off, because each signed project was large.

The contracts did not stop at the end of the month. When I wrote up campaign 1, four prospects had signed and seven were listed as ready to sign.

From what I had seen with other clients, I expected about two more contracts over the following 9 to 12 months from calls held that May. How the booked calls are handled decides whether those later contracts arrive.

Six months adds up. Campaign 4 booked 488 sales calls in 6 months, roughly 81 a month. Its published snapshot is small: 52 of 84 connection requests accepted (61.9%), and 40 of those 52 replied. Treat those rates as an example, not a benchmark.

What has to exist before the first reply

This is the part that takes time and shows nothing on the outside. From my LinkedIn write-up of campaign 3 in the table above: "it does take time to create a B2B client acquisition system that generates great results, usually it's around 30 days."

What gets built in that month:

1. Sending infrastructure. In the same post: "Buy 10-20 domains for cold email outreach and 20-40 email accounts (can be more, especially for big companies)". Never your main company domain.
2. Authentication. SPF, DKIM and DMARC (the records that tell inboxes an email is legitimate, not spoofed) set up for each domain before anything is sent.
3. Warm-up. "Warm up those 20 email addresses for at least 10-14 days". This is the step founders most want to skip.
4. Profiles. The LinkedIn profiles of whoever sends, rewritten to show who you help, how, and with which case studies.
5. Two lists. Separate lead lists for LinkedIn and for cold email, every address verified, with the bounce rate kept under half a percent.
6. Copy and sequences. Messages written per channel, with follow-ups, and case studies inside them.
7. A short video. A video sales letter that a curious prospect can watch before agreeing to a call.

The full technical version, including how many emails each mailbox should send, is in my post on the sending setup: domains, mailboxes and warm-up.

Smirnov Consulting Group is a Prague-based B2B outbound lead generation agency that runs cold email and LinkedIn campaigns for founder-led B2B companies and books qualified sales calls. My team runs this build for clients. Whoever does it for you, plan for about a month where money goes out before the calendar fills.

Questions founders of dev shops ask

How long before a software development company signs contracts from outbound?

Plan on about a month to build the system (lists, profiles, domains, warm-up) before it produces calls steadily. LinkedIn can usually start before cold email is warmed up.

In the campaigns above, contracts were then signed inside the first month of results: 4 in one German campaign, 5 in 30 days in another. More can follow from the same calls over the next 9 to 12 months.

Do referrals still work for software development companies?

Yes, and nothing here replaces them. They are the warmest leads you will get and the best record of who buys from you. You just cannot schedule them, so they work best next to a channel you control.

Which job titles should a software development company target?

Start from who signed your past projects. In IT and software work the signer often sits in IT leadership: a VP of IT strategy, the head or director of IT, or a CTO, and the right one changes with company size. Keep the list narrow: one title, one company size, one country.

Is LinkedIn or cold email better for winning custom software projects?

Both have produced signed contracts above. One ran on LinkedIn alone and signed 5 contracts in 30 days. In another, cold email booked 21 calls from 5,000 people, and LinkedIn with email follow-ups booked 13 from 4,000. We usually open on LinkedIn and add email behind it, then scale whichever channel works for that niche.