Late last year, over on the public defender‘s blog, Pierre Lapree penned a post that asked how many decimals of π does Procurement really need? In short, the answer was, it depends on the context — some situations don’t need very precise calculations, and others need precision down to 1/100th of a decimal point. In his post, Pierre notes that in some situations, like savings, rounding can be to the closest power of 10. In others, like RFP, rounding to the dollar is more than enough, sometimes the closest hundred or thousand is enough. But in spend analysis, sometimes you need to match those financial statements down to the penny to get it right.
But how do you do that?
Your data is a mess, across multiple systems, in multiple formats, with varying levels of detail.
The financial reports are typically created from spreadsheets, which, even though they were output from the organization’s accounting systems, are typically riddled with errors.
And any hopes of matching, despite the fact that each system should be the checksum for the other, are as fantastical as J.K. Rowling’s beasts.
So how do you get precise?
You get out of the data and into the real world. When you don’t know where you are in the real-world, you geo-locate. How do you do that? In today’s world, you tri-angulate your position by taking measurements with respect to cell phone towers and/or satellites and using mathematics to estimate your position as close as possible – the more readings, the more accuracy.
In other words, you take measurements. Lots of measurements. And correlate them. The financial statements are just one set of checks.
Another set of checks are inventory levels. You’re paying for physical goods — you should have payments and invoices for the majority of physical goods and vice versa.
A third set of checks is the accounts receivable system — every part or good that was bought for (re)sale should not only have a corresponding inventory entry but an invoice, and vice versa.
In other words, every enterprise system that tracks goods and services is a data point for correlation, and should be used as such. Don’t just focus on the dollars and cents, as trying to balance erroneous totals can lead you down the wrong path — use all the data at your disposal to get it right — and precise.