DSO Calculation Methods Best Practices

I must agree with many of the comments by Bryan - especially the twitchy ones:).
Perhaps a refinement of his idea is to see if you can gather two sets of data consistently: Actual Billings per month (NOT revenue recognized), and Month-end totals for Unbilled Amounts on Signed Orders.
My reasoning: 1. First metric:calculate DSO on AR and Actual Billings. Use the reduction method where each month's billings = 30 days of AR.
2. Track the trend on Unbilled Amounts.
Then compare this with (a) the DSO above and (b) the Monthly total of new orders booked.
That way you can show your execs:
1. DSO efficiency-cash collection efficiency
2. Unbilled amounts - billing efficiency on signed deals
3. New business booked-sales efficiency in signing new deals

Out of that you may be able to see if any processes need improving:)

Brett, Do you know if there was any reason why your company chose to use this measurement for collection performance? Different compamnies may use different methodology depending on their access to measurement data and other preferences, for example, analysts may look at DSO based on quarterly average days sales if that information is publicly available, whereas internal data may be used to calculate DSO based on lifo or monthly average days sales.

The important thing is that users understand how the data is calculated, what it means, and most important, are able to monitor and explain changes in the trends.

If your revenue is unbilled because you have not "earned" it, then your results will really be skewed. Then your contract to recognized revenue to billed revenue delays can cause big swings in addition to only billing at month end. From a cash flow perspective, are you able to bill throughout the month?

Brett,

Although DSO is a widely accepted metric used in assessing collection effectiveness/capability/performance (as noted above in a few responses), it merely reveals how long it is taking for your customers to pay you. From there, you must employ additional metrics to determine if a high DSO is perhaps caused by (a) pressures/circumstances your customers are facing (economy?), (b) your own internal collection performance (most widely associated issue), (c) some other factor(s) and (d) some combo of the above.

So my suggestion is to think beyond what your eventual DSO calc reveals... don't just conclude that a high number means your team is doing a bad job.

If you get an unacceptably high number, yes, I agree, you should first measure what you control: your collection team and your collection process/policies. So implement additional metrics that show your performance there. For example, you can take the AR Trial Balance and compare the frequency of calls the team makes to the 'Over 30' bucket vs. the 'Over 60" vs. the 'Over 90', etc. Assuming you have a Collection Department policy requiring X number of calls per month to each bucket, and that those calls are documented, you can create at least two derivative metrics to assess performance: (1) % of accounts called per bucket and (2) collection agents in compliance with the call policy.

Additional metrics to consider are: your invoicing accuracy, billing frequency (as you/others mention above), policy on timing to engage external collection agency, etc.

There are many things in addition to those already listed.

1. include or exclude intercompany sales, depending on what is desired
2. exclude items in A/R but were unusual items which may not have run through sales, e.g. you may bill for co-op advertising (manufacturer assists a franchisee/dealer with advertising) or other items. What are the things that are included in the numerator but not in the denominator and vice versa
3. exclude sales allowance accruals, the actual amount billed should be used, not "net sales" from the P/L
4. is this a "month end" only calculation or is it practical to calculate more frequently? Depending on your billing and payment due dates the DSO could misrepresent reality if the billing and terms cause payments to always come in the first few days of the month
5. What you are able to do may be restricted by the capabilities of your IT system.
6. Ultimately, in my opinion, don't worry so much about the DSO number and focus on the individual accounts past due and the DSO will take care of itself.

1. If you are a public company, then I would make sure your method of DSO calculation aligns with other public companies in your industry. This way analysts can have an apples to apples comparison.
2. If the DSO calculation is for internal reporting only, the key is to pick a method(s) and stick with it over time to be able to spot any trends.

If you don't have access to internal data, DSO is the best measure available. However, I would NEVER use DSO if you have access to internal data.

The major problem with DSO is that it's strongly influenced by the growth rate of your sales. This is a problem that can't be solved by finding a "good" averaging period, or by using a consistent calculation method.

Another problem is that DSO can be influenced by a change in the mix of sales made under different credit terms.

Still another problem is that the calculation produces a measure that is too gross. Instead, if you want to improve the time between shipment and collection, you need visibility into the various delays in the chain of events that connect those two end points.

The solution is to directly calculate the length of time it takes for each delay in the collection process: from shipping to billing, to transmission of the invoice, to the customers' payment approval, to check generation, to mailing, to depositing the check, and so on.

Most of those steps leave a data trail that can be captured and analyzed using a variety of statistical techniques, from simple to sophisticated. SMB-type companies can use Excel, as suggested in the article. Larger companies have additional choices.

But NO company should use the DSO for internal control. It tells you nothing useful.

At the risk of re-covering what has already been said above, I have two overriding comments:
1. DSO should always be calculated using the value of invoices posted in a given month and not what was recognized as revenue. Think of it from your customer's perspective. He/she is being asked to pay a specific amount, based on specific terms provided. That customer neither knows or cares what your organization's policy is regarding revenue recognition. The fact that the customer is, for example, paying in advance for a quarter, is all he/she knows. If you recognize that revenue evenly over the three months, you're still expecting payment of the full amount inside the terms given.
2. Following on from the above, average DSO is not able to recognize seasonal fluctuations. Using the quarterly payment above, your cash flow projections are going to be distorted if you apply an average DSO to every month's sales, when in fact the sales lag the amount invoiced by 30 and 60 days in every quarter.

If it's of any help, I've developed a very comprehensive, Excel based forecasting and analytics tool, that has an Accounts Receivable module. This module uses the lookup function to go back and add up all of the individual monthly invoice totals, up to a maximum of 12 months, to calculate an exact AR balance based on the DSO provided. If you think this can help you let me know. My email address is [email protected]

Brett,

People often use DSO to forecast collections in a cash flow forecast. Because you're a Controller, this might be a topic that interests you.

Think of a standard AR aging schedule with buckets for 0-30 days, 31-60 days, and so on.

--In January, the 0-30 day bucket shows the amount of sales billed in January that have not yet been paid in the month of the sale.

--In February, the 31-60 day bucket shows the amount of sales billed in January that have not yet been paid 30 days after the month of the sale.

--In March, the 61-90 day bucket shows the amount of sales billed in January that have not been paid 60 days after the month of the sale.

--Etc.

So if you arrange these numbers in a spreadsheet, you can start to see a pattern, perhaps like: about 5% of your billed sales are collected in the month of the sale, 50% is collected in the second month, 30% in the third month, and so on. And some (hopefully small) percentage of these sales never will be collected.

This collection pattern is very useful in cash flow forecasts when you convert forecasted sales into cash collections. I've found it to be a much more accurate method than merely using a formula based on the DSO.

Also, if your collection averages tend to vary significantly, you can vary the percentages in your forecast to give you a fairly accurate best-case / worst-case collections forecast. (There could be a seasonal pattern in these collection percentages, by the way.)

In contrast, the simple-minded method of scaling up recent sales by the DSO to forecast the AR balance tends to be much-less accurate...particularly if you forecast monthly sales to vary significantly from month to month.

I would attack this using two calculations. Before doing this, you should eliminate any thought of doing a “daily count back” DSO calculation as you do not have sales data for every day.

Your first calculation should be based on revenue. This is more encompassing of your total available cash collection “basket.” Even though, some of the other responders tell you not to do this, it will tell you how good or bad your collections are against the revenue you have recognized. I am sure you and your managers want to know how well you are doing against your “potential to collect.” This calculation would include unbilled revenue and or billed, but not recognized. The simple calculation would be taking you’re A/R balance and extinguishing it going backwards in time. For example, if you’re A/R (including unbilled) is $800 and revenue was $500 (30 days) in the most recent month and $700 (31 days) prior to that, DSO would be 43 days. 30 days + ((800-500)/700) X 31 days or 30 +13=43 days

If you just want to measure your collections on invoices billed, use only your billed dollars in the revenue calculation above and the accounts receivable balance (excluding unbilled/billed not recognized).

The first calculation would be what I would use for external reporting and for all levels of management. The second calculation is what I would use in my collectors performance reviews since this is what they can control. We only use the first calculation throughout the organization as our billing is done daily. If done properly, you may see more activity where you would have more billed-not recognized receivables. That would be the behavior to strive for. Eventually, you can eliminate one of these calculations based on what your organization favors.

I agree with many of the comments above, especially Patrick's. We used DSO, as calculated above, to assess level and trend for industry benchmarking purposes. However, we used the collection efficiency ratio (% collected out of available to collect) at the individual collector level. That ratio was well respected at the National Association of Credit Management.

Everyone, I am just a credit manager, but for me, tracking and graphing my incoming daily/weekly receivables as a percentage against my outstanding AR @ month-end is a great way to understand how my customers are performing and if I'm on target in collections. Once you track this, you'll see that you typically have a trend by week. Another item I like to track is how much my receivables roll as a percentage of their original balance...for example, after 30 days what percentage of the once current amounts rolled into 30 day - how much of that original ends up in 60 days? Once again, if you track over time, you know your trend and see what the true performance is... Sales go up and down, but tracking just your billed receivables gives a truer picture of your AR performance.

Wow - this topic drew lots of responses. You may not be public now and you may not be seeking outside investment now but that may change. I would use the metric most often used in your industry and calculate it based on how an outsider would do it. Use the AR total that appears on your balance sheet and use the revenue as it appears on your P&L. If you do not do that you will find yourself constantly reconciling to those numbers. Ideally take your quarterly revenue and divide by 90 to get revenue per day. Then divide that into AR to get DSO. Note that this includes unbilled AR and it should because the offset to unbilled AR is revenue.