Cash Flow Forecasting From CRM Data

To forecast cash from CRM data you need one more date on every deal: expected cash date = close date + delivery time + payment terms + average collection delay. A sales forecast answers "when will this be signed". A cash forecast answers "which week does the money land". Companies that treat the two as the same thing hit their number and still run short in payroll week.

Why is a cash forecast not a sales forecast?

A concrete example. In September you sign a $120,000 deal. Delivery takes 45 days, you invoice on delivery, terms are 60 days, and this customer runs about 18 days late. The money arrives not in September but in early February — roughly 120 days after signature. Your sales report celebrates September while your bank balance struggles in December.

As that gap widens, companies become profitable and cash-poor at the same time. It is the most common financial accident in a growth phase: as sales rise, the costs you pay up front rise with them, while collections arrive on the same old delay.

What data does a cash forecast need?

Deal amount and close date: the two basic CRM fields.
Win probability: by stage, from historical data rather than a guess.
Delivery or service time: days from signature to invoice.
Payment terms: per customer; one blended average will mislead you on large accounts.
Average delay: how many days past terms the customer actually pays, taken from collection records.

The last two usually do not exist in the CRM — they live in accounting. The real work of building a cash forecast is bringing those two fields into the CRM.

How do you build it step by step?

Lay out open deals by month according to close date.
Weight each deal by its stage probability (for example: quote 40%, verbal yes 75%, in contracting 90%).
Add delivery time, payment terms and customer delay to each row and compute the expected cash week.
Book the weighted amount into that week; if there is a deposit, split it (say 30% on signature, 70% on delivery).
Layer existing receivables on top — the most certain cash line you have.
Put the resulting weekly table side by side with your fixed-cost calendar.

The output should be a 13-week chart, not a single number. Thirteen weeks covers a quarter plus the first month of the next one — enough runway to see a squeeze before it arrives.

How do you set the probability weights?

From historical conversion, never from a rep's instinct. If 38% of deals that reached the quote stage in the last 12 months were won, that stage is weighted 40%. Ask a rep for a percentage and the average answer is 70%; actual results usually come in at half of that. In cash forecasting, the cost of optimism comes back as interest on a credit line.

How do you calculate average delay?

Take the last 12 months of collections, compute (payment date − due date) for every invoice, and average it per customer. In most companies the table splits cleanly: some customers run 3-5 days late, others 30-45. Treating both groups as equal in a forecast breaks it from the start. Shortening terms or asking for a deposit when selling to the second group is the fastest way to repair a cash forecast.

How do you build the scenarios?

Three scenarios are enough, each moving one variable:

Good: weights as they are, delay at the historical average.
Middle: weights 20% lower, delay 10 days longer.
Bad: the two largest deals slip a quarter, delay rises by 20 days.

Whichever week goes negative in the bad case, you need your credit line arranged at least six weeks before that date. The point of a cash forecast is not to know the future — it is to make the decision early.

What do you measure, and how often?

Weekly: the 13-week chart and forecast accuracy (forecast cash vs. actual).
Monthly: realised conversion rates by stage, which is how the weights get updated.
Monthly: average delay per customer.
Quarterly: average days from signature to cash. That single number shows you what growth costs in cash.

What are the most common mistakes?

Taking deal amounts net of tax and mixing them with gross collections on the cash side.
Ignoring deposits — the earliest and most certain line in the whole forecast.
Using one blended payment term for every customer.
Forecasting monthly. Payroll, tax and rent land in the same week; a monthly view hides an intra-month squeeze.
Leaving close dates to reps without verification. A deal carrying a close date already in the past quietly corrupts the forecast.

The quality of this forecast depends on one thing: that the close date, the terms and what the customer actually said are on record. Closync extracts that from the conversations, writes it into the CRM and interprets it across four layers to simulate what comes next — so the cash forecast becomes a continuously updated output rather than a spreadsheet someone rebuilds by hand.