Is DSO really an Accounts Receivable performance KPI?

Search for “key accounts receivable performance KPIs” and one metric almost always appears near the top of every list. Days Sales Outstanding (DSO), which is presented as basic measure of collection performance with simple interpretation. Lower is better. Higher is worse. If DSO rises, then the accounts receivable team is expected to explain what went wrong.

A mainstream view, where benchmarking sites like APQC classifies DSO as a KPI that signals “how well your accounts receivable processes are performing”. Commercial guidance repeats it, usually just before a pitch about how many days their product will shave off your DSO.

Indeed, DSO is a KPI. But my question here is whether it is a performance KPI for the team collecting the cash.

Which question is actually being asked?

Monthly performance review is under way and DSO came in three days above forecast. Attention turns to accounts receivable and the question “what happened?” is raised.

One question on the surface. Two underneath, and they are rarely pulled apart.

First is a performance question. Did the collections team this month do a worse job?

Second is a variance question. Why is reported figure different from the figure we expected?

Most of the time the room wants an answer to the second. However, question is aimed at the team responsible for the first. A miss in the forecast is quietly treated as weaker collection performance, because DSO is the only number on the slide.

If we look at the formula, it converts receivables balance into an approximate count of sales days. It is a legitimate working capital ratio. What it cannot do on its own is tell us whether collections team performed better or worse.

Calculation cannot isolate collection behaviour, because receivables balance can move for reasons that have nothing to do with collections. So can sales figure it’s divided by.

Take payment terms. A customer moved from 30 to 60 day terms can pay exactly on time and still increase DSO. Nothing about collection performance deteriorated, but commercial condition changed.

Sales volume creates same attribution problem. Receivables sitting on balance sheet were generated by sales made across different periods. Conventional DSO relates that balance to sales figure selected for the calculation. When sales pattern changes, ratio can change without any change in underlying collection experience.

Nor can DSO variance be attributed to collections behaviour from DSO figure alone. It reports one figure. No breakdown shows what drove it. Asking AR to account for DSO movement asks team to reverse engineer cause from number that was never designed to carry one.

The objection is fifty years old

Wilbur G. Lewellen and Robert W. Johnson (1972), Lewellen and Robert O. Edmister (1973), James Gentry and Jesus De La Garza (1985), and Gallinger and Ifflander (1986) all address same underlying problem. Conventional receivables measures combine changes in collections with changes arising elsewhere. That blending makes it difficult to isolate actual cause of improvement or deterioration.

Leitch and Lamminmaki returned to the issue in 2011 and put it directly in performance-measurement terms. A change in the conventional average collection period (another name commonly used for DSO) does not necessarily represent a change in collection efficiency. Their proposed refinement again links outstanding receivables back to sales that generated them.

Common direction is decomposition. Separate effect of sales pattern from effect of collection experience, and report each on its own. That gives the investigation its starting point: how much of the movement came from receivables and how much from sales, including where the two offset each other. Which team gets asked what should follow from that result.

Fifty years of published objection, and DSO still sits near the top of those KPI lists. Problem is not simply awareness. Plenty of practitioners know its limitations and report figure anyway.

DSO is obvious candidate for peer benchmarking, and that alone appears to have been enough to preserve it. It can be calculated from published accounts. Most alternatives require additional data, such as invoice-level ageing and contractual due dates, that outsiders cannot see.

Yet benchmark itself is loose. There is no single calculation convention. APQC uses average gross accounts receivable divided by total gross annual sales, and excludes unbilled receivables. AFP uses average accounts receivable divided by net revenue. AFP notes credit sales would be more accurate, and notes many public companies do not disclose split between cash and credit sales. Both calculations are defensible, but they are not same calculation.

Benchmarking body can impose its own convention and create comparability within its dataset. But that is benchmarking body’s DSO. It may not be DSO that company calculates and manages internally.

A shared calculation convention makes comparison possible. Judging collection performance needs more, since two businesses can calculate DSO identically and report different results because they offer different payment terms.

Checklist: is your DSO answering either question?

1. If AR had collected exactly the same invoices on exactly the same days as last month, would this month’s DSO be the same number?
2. Can you state how much of the movement came from the sales pattern and how much from collection behaviour, as two separate figures?
3. Who controls the denominator? If nobody in AR does, what exactly is the metric assessing?
4. If customer terms move from 30 to 60 days and every invoice is still paid exactly on time, has AR performance deteriorated?
5. How much does the figure change if one large customer pays a day either side of period end?
6. Is the target derived from what AR can influence, or from last year’s figure minus two days?

A DSO movement in either direction says little about collection performance on its own. Before it becomes a judgement on collections, the reporting pack should show how receivables and sales contributed to it. Where that analysis points to changed payment behaviour, AR explains what it knows and investigates the accounts concerned.

Part 2 asks whether most monthly DSO movements need an explanation at all, how much of a monthly DSO movement is normal variation, and what a control chart says about which months deserve an explanation.

---

References

APQC, What is DSO in Finance?. https://www.apqc.org/resources/blog/what-dso-finance

APQC, Days sales outstanding, Open Standards Benchmarking. https://www.apqc.org/resources/benchmarking/open-standards-benchmarking/measures/days-sales-outstanding

Association for Financial Professionals (2025), Days Sales Outstanding (DSO). https://www.financialprofessionals.org/glossary/days-sales-outstanding-dso

Lewellen, W. G. and Johnson, R. W. (1972), “Better Way to Monitor Accounts Receivable”, Harvard Business Review, May–June, pp. 101–109

Lewellen, W. G. and Edmister, R. O. (1973), “A General Model for Accounts-Receivable Analysis and Control”, Journal of Financial and Quantitative Analysis, 8(2), pp. 195–206. https://doi.org/10.2307/2330014

Gentry, J. A. and De La Garza, J. M. (1985), “A Generalized Model for Monitoring Accounts Receivable”, Financial Management, 14(4), pp. 28–38. https://www.jstor.org/stable/3665334

Gallinger, G. W. and Ifflander, A. J. (1986), “Monitoring Accounts Receivable Using Variance Analysis”, Financial Management, 15(4), pp. 69–76.

Leitch, P. and Lamminmaki, D. (2011), “Refining Measures to Improve Performance Measurement of the Accounts Receivable Collection Function”, Journal of Applied Management Accounting Research, 9(2), pp. 1–20. http://hdl.handle.net/10072/41011