Your Customers Show Trouble in Their Money Before They Show It to You
Platforms watch behaviour: logins, tickets, payment dates. By the time those move, a business has usually been struggling for a while. A practical guide to reading the numbers underneath, for the people who own the customer relationship.
Read
8 min read
Published On
Oct 7, 2026
Read: 8 minutes
Last updated: October 7, 2026
In this guide
- Behaviour is a lagging signal. The money moves first
- Six cash signals an underwriter watches, and where to find each
- A worked example of a customer that looks fine and is not
- What to say when you reach out, and what never to do with a warning
Behaviour Lags the Money
The story is familiar to anyone in customer success. A customer who logged in daily logs in less. A payment arrives a few days late, then a week late. A ticket asks about cancelling. Then they are gone, and everyone says they should have seen it coming.
They could have. Each step of that story has a counterpart in the business's own cash, and in our view the cash usually moves first. This is a working view, not a measured law, and some businesses fail suddenly. But a platform that waits for behaviour to change is reading the last signal, not the first.
What Your Platform Sees and What the Money Shows
Take Harbor Dental, a three-chair practice on a practice-management platform. Staff are paid every two weeks, 14 people in all. Receivables have always been steady. Every figure below is illustrative.
| Week | What your platform sees | What the money shows |
|---|---|---|
| 1 | Normal logins, subscription paid on time | Deposits start to flatten |
| 4 | Normal logins, subscription paid on time | Deposits down, costs up, balance thinning before payroll |
| 8 | Fewer logins, a payment two days late | Cash covers about nine days of outflows |
| 12 | A ticket asking about cancelling | A missed payroll is likely |
Your window in this picture is weeks 1 to 8. Behaviour alone gives you week 8 at the earliest, and that is already a rescue rather than a favour.
Six Cash Signals Worth Watching
These are the things an underwriter reads in a bank history. Treat the table as a reading guide, not a rule set. You decide your own thresholds, and one odd week is not a trend.
| Signal | Why it matters | Where you might see it |
|---|---|---|
| Days with a negative or near-zero balance | The business is spending cash before it arrives | Bank data the customer has connected |
| Returned or failed payments | A payment bounced, often the first hard sign | Your own payment and debit records |
| Deposit volatility rising | Income is getting less predictable | Deposit history by week |
| Revenue concentrated in one customer or processor | One loss or delay sinks the month | Deposit sources |
| Other advances being added | Repayments are taking a growing share of sales, a pattern called stacking | Recurring debits in bank data |
| Cash cover of the next payroll, in days | A tight week has a date on it | Balance, run schedule and pending debits |
A Worked Example: Harbor Dental in October
Reading Harbor Dental (illustrative)
Six months ago: deposits averaged about $118,000 a month. The balance comfortably covered two payrolls.
Now: deposits are around $96,000 for the third month running, about 19 percent lower. Supply costs are up and a new equipment payment started. The balance covers about nine days of outflows.
What your team knows without the numbers: nothing yet. The practice still logs in, pays on time and has opened no tickets.
That last line is the point. Behaviour looks fine because the problem is still small. A call this week is a favour. A call after a missed payment is a rescue.
A Review You Can Run Across a Portfolio
You cannot call every customer, so rank them and read the few at the top.
- List your customers and attach the cash signals you can see. Even two or three signals sort a portfolio.
- Rank by who is closest to a date that hurts: the next payroll, rent or debit.
- For the top five, write one sentence in your own words about what is happening.
- Decide the action: a check-in call, a note from the account manager, or nothing yet because someone is already talking to them.
- Keep a short log of who you contacted and what happened. After a month you will know which signals your customers act on.
Start with your five largest customers, not all of them. A small review that runs weekly beats a big one that happens once. How to Calculate Cash Runway for a Small Business gives the arithmetic behind the days-of-cover signal, and Why Profitable Businesses Run Out of Cash explains why a healthy income statement can sit beside a thin balance.
What to Say When You Reach Out
Lead with the business, not a product. The owner should feel noticed, not targeted.
Hi Priya, I was looking at how the practice is doing and wanted to check in. The last few months look slower, and supply costs haven't helped. How are you feeling about the next couple of months? Happy to talk through options if that would be useful.
A sample check-in. Adjust to your voice.
It names no amount, makes no promise and mentions no score.
What Not to Do With an Early Warning
- Do not announce a label. Telling an owner your system marked them as slipping reads like a grade. Talk about the numbers.
- Do not turn every flag into a capital pitch. The point of calling early is the relationship. Many flags are a reason to check in and nothing more. If the owner already knows and has a plan, thank them, note it and move on.
- Do not over-read one week. Look at several months before you act.
Sizing the Value for Your Own Base
Multiply the customers you lost last year by annual revenue per customer, then by the share you could plausibly keep with an earlier call. Be careful with the middle step. How many of your losses were cash-driven first is unknown until you measure it, so pull last year's churned customers and read their deposits in the months before they left. Most teams find the story was there.
Where capital fits is a separate question and a smaller one. Slate funds businesses with merchant cash advances and credit lines, through the platforms they already use. It is one option when it fits, never the lead. For the economics of a program, see How Embedded Capital Works for Platforms. To go deeper on a payroll platform's version of this problem, read Payroll Shortfalls: How a Platform Can Spot a Short Week Before the Run.
Slate Signals is on a waitlist. Join it.
Common Questions
Will this catch every struggling customer?
No. You can only read the records you have, and a business that connects little will tell you little. Treat it as a way to find the ones you would otherwise miss.
How do I pick thresholds?
Start from your own history. Look at customers who left and see which signals moved first, then set a level that gives you a list you can actually call. Review it quarterly.
What if the owner asks how we know?
Tell them what you read: their own deposits, costs and balance. Do not describe a score you cannot explain.
Should every flag lead to a capital conversation?
No. Capital is one option when it fits. Many flags are only a reason to check in.
Read next: Payroll Shortfalls: How a Platform Can Spot a Short Week Before the Run · December Makes Your Best Customers Look Healthy. January Is When It Shows. · Why Profitable Businesses Run Out of Cash
