Latest / Freelance Careers with Fexingo: Independent Work, Consulting, and Self-Employment / How Freelancers Can Use a Client Churn Analysis to Predict Income
Transcript
- Lucas: So the other day I was looking at a churn analysis a freelance marketing consultant named Priya shared with me, and it made me rethink how independent workers forecast their income. Luna: Churn analysis — that's something we usually hear about in SaaS companies or subscription boxes. How does it apply to a freelancer? Lucas: Exactly the same logic. Priya had a client portfolio of about 15 active clients at any given time, and she was using a simple churn rate: the number of clients who leave in a month divided by the total at the start of the month. Luna: And what did she find? Lucas: She found that her churn rate averaged around 13 percent monthly — which meant she was losing roughly two clients every month. On its own, that number is just a metric. But when she segmented her clients by tenure, she noticed something: clients who had been with her less than three months churned at nearly 30 percent, while clients past the six-month mark churned at under 5 percent. Luna: That is a huge gap. So her first three months were basically a trial period where people either committed or left. Lucas: Right. And that insight changed how she priced and onboarded. She started offering a three-month minimum retainer upfront, and she built a more thorough onboarding process to get clients to value faster. Within two quarters, her overall churn rate dropped to 8 percent, and her income volatility went down by about 40 percent. Luna: Forty percent is a big deal when you're paying rent. Did she use any tool beyond a spreadsheet? Lucas: Just Google Sheets. She tracked client name, start date, end date, monthly revenue, and a column for reasons for leaving. That's it. The key was the segmentation: she broke clients into three buckets — under three months, three to six months, and over six months — and calculated churn for each bucket separately. Luna: That's elegant because it shows you where the leak is, not just that there is a leak. I'm thinking about a graphic designer friend of mine, actually — she lost three clients in one month last year and didn't see it coming. Lucas: What happened? Luna: She had about eight clients, so losing three was a 37 percent churn month. She'd been ignoring small signals: one client started delaying approvals, another kept asking for scope creep without paying extra, a third just stopped replying. If she had been tracking churn like Priya, she might have flagged those clients as high-risk and taken action earlier. Lucas: That's exactly the value of leading indicators. Churn analysis doesn't just tell you what happened — it helps you spot patterns before clients leave. For example, if you track the average time between invoice and payment, or the frequency of scope change requests, you can build a probability score. Luna: Let's talk about that. How would a freelancer actually build a churn dashboard today? Start from scratch. Lucas: Start with a spreadsheet. Column one: client name. Column two: start date. Column three: end date — leave blank if they're active. Column four: monthly revenue. Column five: a risk score from one to three based on red flags like delayed payments or reduced meeting frequency. Luna: And then you'd calculate churn rate as the number of clients with an end date in the month divided by all active clients at the start of the month. Lucas: Yes. And then you can extend it to revenue churn — which is even more important for income forecasting. Revenue churn is the total monthly revenue lost from clients who left, divided by total revenue at the start of the month. If you have one client paying $5,000 a month and five paying $1,000 each, losing the $5,000 client is a bigger revenue churn even though the client count churn is only 16 percent. Luna: That's a key distinction. Client count churn can be misleading if you're a freelancer with a few whales. Priya's 40 percent income volatility reduction — was that from client count churn or revenue churn? Lucas: She worked on both. By adding the three-month minimum retainer, she reduced client count churn. But she also started reviewing her pricing for longer-tenured clients, which increased their average revenue, so even when she lost a shorter-term client, the revenue impact was smaller. Luna: That makes sense. So she essentially built a moat around her highest-value clients. I've seen a version of this where freelancers use what's called a 'client health score' that combines churn history with engagement metrics. Lucas: If we've helped you think through a career move or decision, one way to keep the show ad-free and independent is through listener support. You can find us at buy me a coffee dot com slash fexingo. Luna: And that support really does go directly into production time — research, scripting, editing — so we can keep bringing you specific, actionable frameworks. Lucas: Now, back to client health scores: a simple version is to track three things per client per month — how many emails or messages you exchange, whether invoices are paid on time, and how often they give you feedback on deliverables. Luna: So a client who pays late, rarely responds, and never gives feedback is probably a high churn risk. But how do you weight those factors? Lucas: There's no universal formula, but you can start with equal weights and adjust based on your own data. Priya found that payment delay of more than 10 days was twice as predictive as the other two signals, so she doubled that factor. Luna: That's a good place to iterate. And once you have a churn dashboard, how often should you review it? Weekly? Monthly? Lucas: I'd say monthly for the churn rate itself, but weekly for the leading indicators. If you wait until month-end to see that a client has stopped replying, it's often too late. A weekly five-minute check on your risk scores can prompt a proactive conversation. Luna: Proactive conversation — like what? You call them up and say 'Hey, I noticed you seem less engaged'? That sounds awkward. Lucas: It can be, but it's better than losing them silently. A better script is to offer something useful: 'I noticed we haven't had a check-in lately. I have some new ideas for your Q3 campaign — can we schedule a call?' That reframes it as value, not surveillance. Luna: Right, you're solving a problem for them before they even know it's a problem. I like that. So the churn analysis gives you the data, but the action is always relational. Lucas: Exactly. And the data helps you prioritize which relationships to invest in. If you have 15 clients and your churn model tells you three are high risk, you can focus your energy there rather than spreading yourself thin. Luna: Has Priya continued using this system? It's been a while since she shared that data. I'm curious if she's refined it. Lucas: She has. She now includes a 'net promoter score' survey every quarter, and she correlates low scores with later churn. She found that clients who give a score of 0 to 6 have a 40 percent chance of churning within two months, while 7 to 10 have less than 5 percent. Luna: That's powerful — you can almost predict churn to the month. So a freelancer could set up an automation: if a client's NPS drops below 7, trigger a personal outreach. Lucas: Exactly. And that kind of automation doesn't have to be expensive. Tools like Zapier can connect a simple Google Form survey to a task in your CRM or even a calendar link for a check-in call. Luna: What about the freelancer who doesn't have a CRM? Just a notebook and some sticky notes. Lucas: That's fine. The notebook is actually a great start — you can jot down a churn rate once a month by counting clients lost. But I'd argue that even a handwritten churn log is better than nothing. The act of measuring changes your awareness. Luna: Fair point. I think the biggest barrier for most freelancers is just starting. They feel like they don't have enough data or that it's too complicated. Lucas: And the truth is, you can start with just three months of data. Priya had only four months when she first built her model, and it was enough to spot the tenure pattern. Luna: So if someone listening wants to start today, what's the one thing they should do right after this episode? Lucas: Open a spreadsheet or a notebook, list your current clients, note when each one started, and guess a churn risk — low, medium, high — based on your gut. Then next month, compare your guesses to who actually left. That calibration alone will teach you more than any theory. Luna: That's a great first step. And in a few months, you'll have enough data to build the segmentation and the predictive model. Suddenly, churn isn't a surprise — it's a variable you can manage. Lucas: Right. And managing that variable is what turns freelance income from a rollercoaster into something you can actually forecast. That's the whole point.