One phrase that's increasingly on the lips of technology vendors selling business accounting and finance solutions these days is 'continuous close.' As an ideal it sounds marvelous — who wouldn't want to always have perfectly up-to-date metrics on their organization's finances? But many must wonder whether it's practically achievable. Traditionally, finance teams have closed their books every month or quarter, in a process that can take days or even weeks to finalize all the adjustments and arrive at the final balances. To any reasonable observer, it's going to seem daunting if not impossible to take all of that work and do it on a daily or continuous basis.
I think the vendors are missing a trick here by focusing on the end goal of continuous close rather than the mechanisms that make it possible. In their enthusiasm to promote the concept, they make it sound harder to achieve than it really is. Continuous close doesn't have to be a sudden-death proposition that you turn on overnight. The point is to deliver more accurate and up-to-date information to the business, without having to wait to formally close the books every month. In reality, many of the ingredients are relatively easy to put in place one step at a time, and it makes a lot of sense to take these steps as soon as you can.
I came to this realization having thought about how I do the books for diginomica's business. We're able to take advantage of a huge amount of automation, which means that we have a pretty good picture of how the business is faring at any particular moment. We only do a formal close once a month, but that only takes about an hour to complete, because virtually all the work has been done on a continuous basis. If my colleagues wanted, it would not take much extra trouble to switch to a weekly or even a daily close, but at the moment it doesn't justify the extra resource this would take.
My takeaway therefore is that many of the mechanisms you need to put in place to enable continuous close are worth doing, even if you're not planning to formally dot the i's and cross the t's every single day. Just getting to a position where you have the option of doing that on any given day should the need arise is already very valuable. And in the process of reaching that position, you're going to remove a lot of manual drudgery and error-prone retrospective adjustments, all of which will make your life much more bearable.
Continuous close means 'shift left'
Fundamentally, the key to enabling continuous close is a principle that software engineers call 'shift left'. To shift left is to move critical activities as early as possible in the lifecycle of the process. In the context of the financial close, that means doing them as close as possible to each individual transaction, rather than applying them retrospectively in a completely separate operation later on. This works best when automation can be put in place to make it easier to complete those activities around the time of the transaction.
This brings us to another technology that vendors are enthusiastically promoting to finance teams, namely artificial intelligence (AI). Once again, I fear they are shooting themselves in the foot by making it sound like something that requires a lot of specialized technical skills. In many cases, adopting AI is simply a matter of plugging in ready-to-run automation tools that simplify mundane processes. There's an increasing amount of AI being applied to my own bookkeeping, but I barely think about it because the vendors have embedded it in the services I use. I don't really care what the underlying technology is, so long as it works. What matters is what the technology makes possible, rather than the technology itself.
I do wonder whether some vendors are making a big deal of AI or continuous close in part because they want to justify a big price ticket for these capabilities. In the past five years, I've often been surprised to sit in a conference keynote or read a press release and discover that some enterprise application vendor is excited about its plans to introduce capabilities I've already long taken for granted in our own finance software at diginomica. For the record, our systems, which cost us around £1,200 ($1,670) annually, comprise Xero for multi-currency accounting and billing, coupled with Dext (formerly Receipt Bank) to process supplier bills, and Flowrev for cost and deferred revenue recognition. This combination provides the following automation capabilities, several of them underpinned by AI:
- Automated import of bank feeds (from a dozen accounts held at four different providers, and in several currencies)
- Automated matching of statement lines to approved transactions in bank reconciliation
- Machine reading of incoming bills and receipts
- Automated coding of bills and receipts from known suppliers
- Automated generation of repeat invoices from templates, including revrec period
- Automated creation and posting of journals for cost and revenue recognition
Moving away from monthly data
Obviously we'd have a lot more complexity if we were a manufacturer, a high-volume retailer or distributor, or had multi-country operations. But the same principles of using automation to 'shift left' apply equally to all of those scenarios. I'll leave the final word to Brian Sommer, who wrote about the need to move away from monthly, aggregated data to more timely and precise data, costs and insights earlier this year. If you want to explore this topic further, the full article is well worth a read. He points out:
Concepts like pay periods, accounting periods, etc. are all arbitrary, man-made inventions that exist because someone decided that we should count or accumulate things at specific, regular times. These determinations were often a compromise where firms balanced out one need (eg: employees would like to get paid in a timely fashion) against another (eg: payroll processing is time consuming and should be done as infrequently as possible). Those decisions, made decades or centuries ago, may have been correct in their day but that day is past.
Behind these timing decisions were manual or automated systems to help with this processing effort. These systems, once placed into production, were often constrained, costly and rigid. So, people and businesses just got used to things being done this way (and frequency) as 'it’s the way we’ve always done it.'
Good news - we don’t have to keep doing things this way!