Latest / Freelance Careers with Fexingo: Independent Work, Consulting, and Self-Employment / How Freelancers Can Use a Cohort Analysis to Predict Income
Transcript
- Lucas: If you're a freelancer, you probably have a decent sense of whether this month is busy or slow. But do you know which months next year are going to be slow — before they happen? Luna: I mean, I have a gut feeling. January is usually quiet because clients are still budgeting. August can be a ghost town. But 'gut feeling' isn't exactly a forecast. Lucas: Right. And gut feeling doesn't help you decide whether to take a discount project in March or hold out for better work in April. So today I want to talk about a tool that startups use all the time, but freelancers almost never touch — cohort analysis. Luna: Cohort analysis. That sounds like something a VC would say about a SaaS company. How does that apply to a solo graphic designer? Lucas: Fair question. The basic idea is simple: instead of looking at all your income lumped together, you group clients by the month or quarter they first hired you. Then you track how much revenue each group — each cohort — generates over time. Luna: Okay, so you're basically saying: treat every new client as part of a class, like a freshman class in college, and see how much they spend over four years. Lucas: Exactly. And what you start to see are patterns. I was talking to a freelance brand strategist in Portland — she does identity work for B2B tech companies. She'd been freelancing for about three years and felt like her income was unpredictable, but she couldn't pinpoint why. Luna: So she ran a cohort analysis? Lucas: She did. She went back through her invoicing software and grouped every new client by the quarter they started. Twelve quarterly cohorts over three years. Then she calculated the total revenue each cohort had generated up to the present. What she found was striking. Luna: I'm guessing Q4 clients were the worst? Lucas: Bingo. Clients who first hired her in Q4 — October through December — had an average lifetime value of about $4,200. But clients from Q2 — April through June — averaged nearly $6,000. That's a 43 percent difference. And the Q2 clients also stuck around longer — an average of 14 months of repeat work versus 8 months for Q4 clients. Luna: That's huge. But why? Is it seasonal budget stuff? Lucas: Partly. She dug into the notes. Clients who came in during Q4 were often in scramble mode — they had leftover budget they needed to spend before year-end, or they were rushing to launch something for January. So the projects were smaller, more transactional, and less likely to lead to ongoing work. Q2 clients, on the other hand, were planning their annual strategy. They had bigger budgets, longer timelines, and a higher chance of repeat engagement. Luna: So the lesson is: don't take on Q4 clients? That seems extreme. Sometimes you need the cash. Lucas: No, the lesson isn't to reject Q4 work. It's to know what you're getting into. If you know Q4 cohorts have lower lifetime value, you can adjust your pricing or your scope. Maybe you charge a premium for those rushed projects. Or you set expectations upfront that the engagement is a sprint, not a marathon. You make a conscious decision instead of just hoping for the best. Luna: That's a much healthier take. So how does a freelancer actually build this? I'm picturing a giant spreadsheet with formulas that take hours. Lucas: It's actually not that bad. You can do it in Google Sheets in under an hour. You need three columns: client name, the date of their first invoice, and the total revenue they've paid you to date. Then you just bucket each client into a quarter based on that start date. Then average the total revenue per cohort. That's the skeleton. Luna: And then what? You look at the averages and say 'Q2 is good, Q4 is bad'? Lucas: That's the start. But you can go deeper. You can also track how many clients from each cohort are still active after six months, twelve months. That retention rate is a powerful predictor of future income. If your Q2 cohorts have a 60 percent retention rate at 12 months, and your Q4 cohorts have 30 percent, you can model what your pipeline will look like a year from now. Luna: So you could forecast your income — not just guess. Lucas: Exactly. And that's the real power. Most freelancers live in a reactive mode. They take whatever comes in. Cohort analysis lets you see the demand cycles in your own business. You might discover that your best clients come from a specific referral source, or a specific month. Then you can double down on marketing during those periods. Luna: Let me push back a little. This sounds great for someone who's been freelancing for a few years and has a decent client history. But what about someone who's just starting out? They might have six months of data. Can they still use this? Lucas: They can, but the signal is weaker. With fewer clients, a single big project can skew the average. I'd say you need at least four or five clients per cohort to get a meaningful number. If you're new, you might group by month instead of quarter, or just track it as a leading indicator — watch how your cohorts are trending, even if the sample is small. Luna: That makes sense. And I suppose you can combine this with other data — like client industry or project type — to get even more specific. Lucas: Absolutely. The designer I mentioned also tagged each client by industry. She found that fintech clients had a 50 percent higher lifetime value than e-commerce clients, regardless of cohort. So she started focusing her marketing on fintech. That one insight changed her business direction. Luna: That's a real strategic shift. You're not just a freelancer anymore — you're running a business with data. Lucas: Right. And that's the mindset we try to encourage on this show. Which brings me to something I want to mention. We keep this podcast free and ad-free — no sponsors, no commercials interrupting the conversation. The reason we can do that is that some listeners choose to support us directly. Luna: Yeah, and we really appreciate that. If you've found an episode useful, whether it's this one or a past one, and you want to help keep it going, there's a simple way. Lucas: It's buy me a coffee dot com slash fexingo. That's it. No tiers, no rewards, just a way to say 'this is valuable to me.' We don't run ads, and we never will. Your support keeps it that way. Luna: And honestly, it makes a difference. Even a small contribution helps us cover hosting and research time. So if that feels right to you, the link is there. And now back to cohort analysis. Lucas: So once you have your cohort data, the next step is to use it to make decisions. Let's talk about three specific actions you can take. Luna: I'm listening. Lucas: First, price differentiation. If you know that clients from a certain cohort tend to be lower value, you can raise your rates for that group. Not in a punitive way — you just adjust your minimum project fee. For example, if Q4 clients usually spend $4,000, set your minimum engagement at $5,000 for new Q4 inquiries. You might lose some, but the ones who stay will be more profitable. Luna: That's a filter. You're using price to pre-select for better-fit clients. Lucas: Exactly. Second, you can time your marketing spend. If your highest-value cohort comes from Q1 referrals, that's when you should be networking, sending out proposals, maybe even running a small ad campaign. Don't blast the same effort across all months. Luna: So aligning your energy with the data. That saves time and money. Lucas: Third, you can manage your cash flow. If you know that Q4 cohorts have lower retention, you can plan for a dip in repeat income during the following summer. Maybe you set aside a reserve in Q2 to cover the slower months. Instead of being surprised by a quiet July, you've already accounted for it. Luna: That's the kind of stability most freelancers dream about. Not having to panic when the inbox goes quiet. Lucas: Right. And it's not about eliminating uncertainty — it's about reducing it to something you can work with. Cohort analysis gives you a historical baseline. You'll never predict perfectly, but you can be a lot less reactive. Luna: So what's the first step for someone listening right now who wants to try this? Should they pull up their invoicing tool today? Lucas: I'd say start with a simple export. Most invoicing tools let you download a CSV of all invoices with client names and dates. Then just add a column for total revenue per client. If you have more than 20 clients, you'll start seeing patterns. Group by quarter, and look at the averages. Even that raw cut will tell you something. Luna: And if the pattern is noisy? Like, no clear Q2 spike? Lucas: Then try grouping by referral source instead. Or by project type. The method is flexible. The goal is to find any grouping that reveals a meaningful difference. If nothing stands out, that's also useful — it means your income is more evenly distributed, and you don't need to worry about seasonal dips. Luna: That's a good point. Sometimes the insight is that there's no problem to solve. Lucas: Exactly. But you only know that after you've looked. And that's the whole philosophy: move from guessing to knowing, even a little bit. Your freelance business deserves the same analytical rigor you'd apply to a client's project. Luna: I like that. Treat your own business like a client. Run the numbers. Make decisions based on evidence. Lucas: That's the takeaway. So if you're a freelancer listening, try a cohort analysis this month. It might just change how you plan your year.