Why operational data doesn’t tell the whole story, and how it ruins your weekend

From: CFO

To: VP of Finance (you!)

Subject: Urgent

Time: Friday afternoon, Sep 30, 2022

The CEO called a meeting to understand why her daily cash dashboard took a turn for the worse this morning. Suddenly our September results are at risk. Can you pull together a deck of supporting data to show her what’s going on?

The meeting is Monday at 9am, so I’ll need the deck by Sunday noon. Sorry.

How many times have you gotten an email like this? Bad enough that your weekend is shot. What’s worse is you know that the deck you pull together will have a bunch of inconsistencies that will drive executives crazy and you need to explain to avoid looking foolish.

Why is it that the CEO’s daily dashboard never seems to tie out when you pull supporting data?

Let’s start by flipping the question, and looking at when the data does tie out, and what it takes to make that happen.

After a month or quarter is over, accounting does a mountain of work in just a few days to pull dozens of reports from lots of different systems and reconcile and adjust the data.

Let’s say you’re a marketplace company that allows individuals to buy and sell, and you also have a separate business that hosts inventory for wholesalers to sell on your platform.

So you’ve got to pull order data from both systems.

On top of that, you offer frequent and new customers promo codes that act as credits on a customer’s account that can get applied against purchases. These credits can be issued when something goes wrong, or as part of a new customer promotion or loyalty program. They can also expire if not used.

So to tie out net order amounts, you have to pull 8 separate reports:

In your person-to-person marketplace:

1. All the orders processed
2. All the credits issued
3. All the credits used
4. All the credits expired

…and the same 4 reports from the wholesale marketplace

On top of that, you use several payment service providers (PSPs), like Stripe, Adyen, and Paypal, and you need to pull reports from each of those services to see amounts invoiced and amounts paid. But you discover that there was an outage that prevented you from pulling one of the reports immediately, so you have to make an estimate and then adjust it next month to stay on schedule.

Once you have these 11 different reports, you have to reconcile them to figure out how much revenue actually came in, how much is waiting to be collected in the dunning process, what the balance in your credits is, and so forth. It’s a huge, complicated dance to figure out the actual revenue and cash balances for the month.

Now let’s go back to the Friday afternoon fire drill.

There are many possible reasons the daily order data might not look great:

• Order volumes in one of your businesses have an issue. This is perhaps the simplest explanation from a financial point of view, though it raises a bunch of operational questions.
• Credits: Perhaps more credits were used than is typical, reducing payments, or more were issued than usual, pulling cash into an account to be drawn against when they’re used. Or maybe fewer credits expired than normal.
• Payment processing: perhaps there was a delay in processing some payments because of an outage in the connection with a given vendor, or the sales mix shifted to a geography with slower payment processing / longer dunning cycles
• An increase in credit card processing fees that reduced cash transfers for a given sales level
• Data issues: Perhaps where a payment processor or order system API changed and the wrong or incomplete data is getting pulled into the CEO’s dashboard. Or you had an outage and none of the orders from Thursday — the last day of the month — were written to the database Thursday night night, so the numbers look off on the CEO’s dashboard this morning. The system will catch up tonight, but by then we’ll have pulled the September data without realizing the issue, and our September results will look soft (and worse, for the controller, when he pulls the data again later, the total will be different and he won’t know why).

To know which of these issues (and it could be one, some, or all of them) is the issue, you have to almost run part of the close process over the weekend. In addition, you have to check with engineering to see if data issues were part of the problem. And no matter what you find, someone is going to catch a lot of heat on Monday, so you’d better be right.

So let’s go back to our original question: why is the daily dashboard not an accurate reflection of the business?

At the root of the problem is that the operational data doesn’t tell the whole story. There’s no fully connected system that ties all the operational data from production systems together and respects accounting rules, so there is no traceability between accounting and transaction systems.

The only time most businesses get the whole story is when they close the books, which, because it’s so complex, only happens once a month.

How to avoid a lifetime of ruined weekends: to avoid this situation, smart businesses don’t let operational data and finance data get too far apart. They use a finance data platform (we happened to think Leapfin’s is pretty good!) to continuously tie out operational data from production systems into reliable, GaaP-compliant financials at a transaction-by-transaction level. That way, when the CEO (or anyone else) asks the question, it’s easy to work backwards to identify exactly what’s changed and why in just a few minutes.