Part of our guide hub: Digital marketing
Software houses tend to grow on referrals, then hit a ceiling when referrals plateau. What follows is the honest performance of each channel, based on what we see across our own pipeline and our clients’. Including the ones that do not work as advertised.
The number that matters
Not leads. Qualified leads: right budget, right authority, real timeline. A channel producing forty enquiries a month of which two are qualified is worse than one producing five of which three are. Most agencies measure the wrong number and optimise themselves into a worse pipeline.
Referrals and past clients
Best cost per qualified lead. Shortest cycle. Highest close rate. And almost always under-worked, because it feels passive.
It is not passive. A quarterly check-in with every past client, an explicit ask for introductions when you deliver something well, and a simple partner arrangement with agencies adjacent to you (design studios, marketing agencies, accounting firms with software-hungry clients) turn a trickle into a channel. If you do one thing from this article, systematise this.
Organic search
Slow to start, compounds indefinitely, best long-term economics. Six to twelve months before meaningful volume, which is why most agencies abandon it at month four.
What works is not general content. It is pages matching the queries of someone already looking to buy: the service page for each thing you build, industry pages describing work in that sector, and articles answering the questions asked immediately before a purchase: what it costs, how long it takes, how to choose a vendor. An article ranking for “ERP implementation cost” attracts a buyer. One ranking for “what is ERP” attracts a student.
Case studies
Not a channel, but the multiplier on every other channel. A buyer choosing a development partner is managing risk, and the only real evidence is that you did this before, in their industry, and it worked.
Three deep case studies (problem, constraints, what you built, measured outcome, honest obstacles) beat twenty logos. The honesty matters: a case study admitting a difficulty is far more credible than one that reads as marketing.
Works well for founder-led content, poorly for company pages. Buyers follow people. A founder or technical lead posting genuine detail (an architecture decision, a project post-mortem, a real opinion about a technology) builds a pipeline over months. A company page posting festival greetings and stock imagery builds nothing.
Outbound on LinkedIn still works, at low volume and high personalisation. Sequences sent to a thousand people do not.
Cold email
Viable but narrow. It works when you have a specific, evidence-backed reason to contact this company: you built the same thing for their competitor, or you can see something concrete about their situation. It fails when it is a generic capability pitch, which is what most of it is.
Deliverability and compliance are real constraints, and a burned domain is expensive. Treat it as precision work, not volume.
Paid search
Fastest to switch on, highest cost per lead, entirely dependent on your landing page. High-intent keywords in this market are expensive and competitive. It works when you bid on narrow, specific intent (a named technology plus a location, or a specific product build), and send traffic to a page about exactly that with a real case study on it. It fails on broad terms and a generic contact form.
Use it to test which messages convert before committing to them organically. That is often worth more than the leads.
Marketplaces and directories
Real volume, heavy price competition. Clutch, Upwork and similar can fill a pipeline, but buyers arrive comparing on price and expect it. Useful for early growth and for filling capacity gaps; corrosive if it becomes your primary channel, because it trains you to compete on rate rather than outcome.
Events and communities
Low volume, unusually high quality. Local tech meetups, industry conferences in the sectors you serve, and genuine participation in communities where your buyers are. Slow, hard to attribute, and the source of a disproportionate share of large contracts.
Where to start, in order
- Systematise referrals. Cheapest revenue available. Start this week.
- Write three real case studies. Everything else converts better once these exist.
- Fix service pages and start organic. Slow, compounding, and the only channel that gets cheaper over time.
- Add founder-led LinkedIn. Free, and it compounds alongside search.
- Then test paid: with a specific offer and a page built for it, not your homepage.
Agencies that struggle usually have no channel producing predictable qualified leads and are trying five at once. Agencies that grow steadily have one channel working properly and are adding a second.
Ezitech runs digital marketing and SEO programmes for technology and service businesses. Talk to our team about pipeline.