The session was moderated by Host Analytics CMO Ben Plummer and the panelists included industry experts Paul Hamerman from Forrester Research. Rob Kugel of Ventana Research, and Steve Player of the Beyond Budgeting Round Table. Here’s what they had to say.
Ben: I’d like to introduce Steve Player, Rob Kugel, and Paul Hamerman. Come on up, guys.
So this is our panel of experts. Instead of me trying to introduce them, since they have such long and illustrious backgrounds, I’ll let each one of them take a few minutes and introduce themselves and then we’ll get into how we’re going to do this.
SP: I’m Steve Player, I’m the program director for Beyond Budgeting Round Table. I do a lot of work with Host Analytics, and spoke at last year’s user conference. You guys may have seen me out on the road show, very excited to be back here sharing the message.
To encourage audience interaction in this session, Host Analytics allowed attendees to vote on panel questions via the conference applications. The questions spanned five key subject areas: finance modernization, cloud and EPM, what’s next for EPM, changing finance processes, as well as budgeting and beyond. Here were the questions selected and the panel responses.
Q: On the topic of finance modernization, how can finance departments use technology to be more strategic?
RK: One of the key things that you buy with technology, if you use it properly … it saves a lot of time. It allows you to eliminate the mechanical, repetitive work that you might be doing today and allows you therefore to spend more time doing more useful things. The shorthand I’ve used for many years is, let’s put the “A” back in FP&A, because if you’re spending all of your time as a spreadsheet jockey you’re probably not doing quality analysis. If I’m preparing data, staging data, correcting things … Get rid of all that because you’re using technology to eliminate that problem, then you’ll have more time to do what you’re really qualified to do, which is analysis, thoughtful work, and really thinking about the rest of the organization.
Part of this is just getting rid of spreadsheets. How do you use technology? You find the spreadsheets that people are spending an awful amount of time working on and try to figure out how you automate the processes to eliminate the need for them. How you connect the organization so that nobody is faking something and entering the system manually and putting into another system and creating more work. Try to automate as much of the work as possible.
PH: Let me take a shot at this from a technology angle. One of the pieces I wrote is about the future of business apps. I think this applies a lot to finance and EPM. There’s really five trends impacting how we use technology in business. One is software as a service, or SaaS. I think we all know what that means, it’s much more sustainable way to deploy applications. Another one is flexibility, that these business applications should be designed to be managed by business people with a minimal amount of IT support. The third piece is analytics. The analytics should be real-time, they should be embedded, they should be pervasive throughout the application.
Another piece is integration. Integration is sort of secret sauce that binds the applications together so when you’re in the planning process you want to be able to pump real-time data into the application, that makes the planning process easier and makes it more valuable. The final piece is the user experience, there’s a lot of evolution in the user experience, which makes it not only better but also inherently mobile. A lot of the engineering that’s gone into the engineering at Host Analytics and other applications are really designed to work on any kind of device, anywhere, anytime using a lot of the consumer paradigm that you’re familiar with from your smart phone.
SP: Being strategic means changing something. The biggest problem we have in finance is that we spend too much time looking at the past. All we’re doing that point is validating. You can’t be strategic staring off the back of the boat. You’ve got use this technology, this innovation to change the role – stop doing dumb stuff. Stop doing budget reconciliations based on a budget with assumptions that turned out to be wrong. As soon as you think they’re wrong, quit looking at them. Change, go to a rolling forecast, and start looking forward. You’ve got to change your role.
Paul’s right about continuous real-time improvement, but when you’re thinking about strategy, we’re predicting the future. If your work today doesn’t include predicting what’s going to happen and looking out trying to understand it … then you’re still looking at the past.
And you can’t predict without looking at the drivers. Then we’re looking at physical things. Physical things don’t need to wait till month end. Right now, we’re still in batch mode. The world doesn’t work in batch mode. In finance, we’re still stuck in batch mode. We process by quarter, we process by year – it moves forward. So you’ve got to turn around and start looking and then, as you’re validating data, you need to look upstream and ask not, did we get the sale, but have we got enough leads coming so that the sales are coming. Get out of working in the past and start working in the future to make sure the strategies come together.
Q: On the subject of cloud and EPM, What’s the best way for finance to partner with IT on EPM deployments?
SP: I learned a long time ago that you only partner with people who want to partner with you. First and foremost, you’ve got to decide, is IT somebody we’re going to be partnering with? Or are they going to conclude that we aren’t worth the work? In many cases, the big problem with finance, is they’re putting in an ERP system -they’ve been putting in the last five years – and it’s going to be at least another five more putting it in. There are other priorities and we never get on the list. If that’s your case, move on. Jump in the cloud. If IT will partner with you and talk about strategy, great. If they’ll partner, partner, If they won’t, don’t. You still have to get your work done.
PH: One of the reasons that organizations like finance and HR have put more of their business apps into the cloud, is because IT let probably them down in the process. They needed their applications up to date, they didn’t give them support that they needed because they were working on the billing system or some other operational systems.
By putting your business applications in the cloud, your financial system, you do reduce a lot of the dependency that you have on IT to maintain software and keep it running. You are, in effect, turning that job over to your software vendor. To keep it secure to keep it running, continually update software for you. It does reduce your dependency on IT. At the same time, you still need them for security, provisioning of the apps, maybe the apps work on mobile devices. You need IT for that. You may also need them for integration. Integration is very important, you’ve got to get business information out of your ERP systems, out of your CRM systems, and out if your operational systems to you can plan and report on how the business is performing.
RK: I don’t disagree with either point of view. In particular, if you’re in the cloud you need to be connected to many other systems within your organization. The more complex the IT infrastructure is in your company, the more you’re going to have some kind of arrangement with the IT people.
I understand the long-standing standoff between IT and finance … What we found over the years in our practice is that the most successful finance organizations are the ones that either have their own internal finance IT group, as well as a good relationship with an IT organization generally. Because they have people who understand the needs of finance, but who can then translate that into Klingon so as to not confuse the IT people. You should have some sort of relationship there.
Q: On what’s next for EPM, how can predictive analytics make forecasts more accurate?
PH: A lot of the planning and the forecasting activities that you do are done manually. It’s the process of making decisions for how to spread money over periods, departments, and so forth. What predictive analytics can do is – analyze massive amounts of data and really detect patterns. But you can also use algorithms to create simulations in very sophisticated ways, as well as forecasts. Another technology that is involved in this is machine learning, and you’ll hear a lot more about it going forward. There’s a lot to watch in here, but some of it is actually mature. Think of some of the scientific algorithms that you use for planning and simulation – those have been around for years. There’s a lot of advanced technology involved in terms of consuming massive amounts of data so I think this is just going to get better and better as time goes by.
RK: One of the things that makes predictive analytics useful is if you identify a list of factors that turn out to be drivers of future results. And if you use even relatively simple techniques, you can very often come up with what those drivers are. It’s so much better when you’ve got tons of data to work with, and lots of data science to work with. But even on a fairly simple scale, it helps you dis-aggregate the factors that tend to drive results. You can be making more nuanced forecasts. And you can update these rolling forecasts as new data comes in, to sharpen and improve forecasts. It can allow you to be more accurate. But equally important, when you find that things didn’t turn out as planned, it helps you put the identifying factors that were wrong that drove the situations where results weren’t what you were looking for.
Along those same lines, one of the things you can do is – as you’re watching events unfold – you can track these things. And when these drivers end up being out of tolerance, you’ll have a quicker idea that things aren’t going to turn out the way you expected and then you can begin to make interventions in the business much sooner. Not after the weak quarter or the month has ended, but you’re doing it intra-period and may be saving the day.
SP: I agree with where they’re coming from, but let me approach it from a completely different angle. To me the problem with forecast accuracy comes from the fact that we have a bad definition of what forecast accuracy is supposed to be. If you don’t get the definition right, you won’t get the forecast right. Most people don’t understand that forecast should merely mean, what we think is going to happen. The problem is that, with most people, we forecast what we want to have happen.
There’s a lot of bias in that the wish becomes the target. If you don’t separate that bias – the target is what you’re aiming for. The target is where you want to go. The forecast has to be totally unbiased, it has to be where you think you’re going. The technology works when the forecast isn’t plagued by the bias that is often overlaid by management teams wanting to “wish it up.” Once you separate those two things, then you have a clean way of checking the science and checking the predictive data to see what actually doesn’t fit. If we find out that the drivers are saying we aren’t going to get to where we need to get to, then you change the action plan. But recognize: you’re changing the action plan that distorts the forecast, that distorts the measurement. You have to understand, when you start changing the action plan, things are going to come out different. We have tons of work to do just to redefine what it means to produce a forecast.
PH: I think there’s a trust factor involved with predictive analytics. We’ve got to learn to trust the machine, and it’s going to feel like sitting in a self-driving car for the first time. You don’t trust that car isn’t going to crash so you’re holding on for dear life. But I think over time we understand more and more that we can trust that this is using powerful technology to help make decisions.
SP: Trust is the understanding that it’s going to do what we think it’s going to do – then we can re-calibrate and adjust where we’re headed.
Ben: I think that’s a great analogy. The self-driving car concept is trust in a predictive model that is going to deliver a forecast that’s based on drivers versus where we want to be. It’s not an easy thing to let go and predictive technology is going to have to come a long way to gain users’ trust.
Q: On changing finance processes, how has technology enabled FP&A teams to streamline their processes and become more insightful? Also, how are planning processes evolving in the digital age?
PH: First of all, when thinking about the past, present, and future and the current finance paradigm it’s useful to think of what most systems do. They help you to manage the traditional processes that are governed by accounting rules in terms of how we report what already occurred. The present is about what’s happening right now. Accounting systems don’t tell you what’s happening right now.
You need analytics and real time reporting systems in order to gauge how the business is playing out. You also need to determine what’s going to happen in the future – forecasting and so forth. That’s really describes a set of business applications that we call EPM. These applications are becoming more dynamic in that they have the ability now to analyze more real-time data. So we can turn a lot of traditionally batch and periodic processes into real-time processes, which I think has an immense amount of value to the business in a number of areas. Take the annual performance review. What’s happening with that right now is that it’s moving to continuous coaching and analytical style performance review and I think that’s going to play out very quickly. Another area is that the processes are becoming more pervasive across the business and involving stakeholders across the entire business. Processes are becoming more collaborative. Today, in the demos, you saw that there is collaboration stream embedded in the application so you can communicate real-time with your colleagues rather than send spreadsheets and get answers back the next day.
Another thing that I think we’ll see is processes become aligned with what people are working on. One of my areas of coverage is human resources and I’ve seen some carryover between EPM and human resources. One of the things we need to do is really match performance at a business level what people are actually working on. And not do it in a cascading way, but really from a modern process where we can align people working on things that really matter to the business.
SP: If you’re looking for leading indicators that begin to predict things like what’s our customer satisfaction? What is their reaction? If we have dissatisfied customers and we’re issuing rebates that’s going to predict that we have issues coming.
If we have satisfied customers, if you have an extra $100 million (like in the demo), the real question is can you deploy that? Again, looking at the past isn’t helpful – what I need to be doing is looking at the future and finding the best place to use that money. What’s going to get us the highest rate of return? It’s really that FP&A is turning around, and the reason it’s happening is the digitization. Think about the Fitbit from the demo. The thing about this thing is, it’s easy. You just click on it, and you already look at it every once in a while. I don’t have to do anything. Sometimes it’ll buzz me and say, here, look at me!
Everything is becoming digitized. The phone in your pocket has more computing power than it took to land a rocket ship on the moon and bring people home safely. That kind of thing – the fact that it’s all digital now – gives us so many more opportunities that we just haven’t figured out how to capture and rethink our processes. It’s going to be about letting go of old processes; we can’t be using this technology to do what we used to do faster. Because then we’ll get to a bad place faster. We have to re-conceptualize what we’re really trying to do – and then use the technology to get there faster.
RK: What else could I add to that? Let me underscore the value of in-context collaboration, which is being added more and more into our applications. And it’s a very important feature. Those capabilities help really streamline the interactions people need to have.
But let me also jump off and say that we have plenty of great technology. The problem isn’t technology. It’s, let me echo a point that Steve made, you’ve got to want these new technologies. In our work we run into a lot of companies with great technology, they just haven’t set it up right. We do operate in batch mode; increasingly that’s not going to be the case since technology is moving that forward. We should still be looking at where are the hang-ups in the systems that you use. Where is the data? How smoothly does the data move in your organization? How accurate is it? There are technologies that are available that allow you to streamline processes. They’re really critical, but no one pays attention to them. If you’re finding that, for example, you’ve got a process and something comes out of one system and somebody’s doing something to it and then handing it back into another system – that’s an opportunity. New technology is going to automate that hand-off. A lot of the problems that financial organizations have is taking care of the little stuff, using technology.
Q: On budgeting and beyond, how can we find time for rolling forecasts when we are struggling to complete normal processing?
SP: You’ve just got to wake up and say I’m not going to go to work and do dumb stuff. Ask why. This isn’t useful anymore. Why are we continuing to do this? Why do we keep closing the books on a monthly basis.
You look at finance, and there’s a lot of stuff that we do because that’s how we’ve always done it. At the end of the day, we’re staring off the back of the boat. If I free up that time, what could I do that would add value? What kind of investments could we make that help us grow? It’s a bold move to say this stuff doesn’t make sense because it’s the accountant’s job.
But fundamentally, it was a job based on a system founded in the 1920s. I think technology has come up with better ways to do this stuff. And when you pull these processes apart and find out what it’s really for, you see all sorts of conflicts of interest so that you can’t get there. What do you want your forecasts to be? I don’t want wishful forecasts, I want realistic forecasts. What do you want your resource allocation to be? I want it to be real-time, I don’t want to think about what the economy is going to be, I’ll over-allocate in some areas and under-allocate in others. Either way, I’m going to have this planning module where I’ve over-allocated.
The whole thing is just dysfunctional. Do we really have to do it the old way anymore? Don’t get me wrong; we don’t need to shrink finance any more. We’ve shrunk finance from about 3% of revenue to a half-percent of revenue. We’ve shrunk finance, but we haven’t transformed it. And there are far better ways to hold people accountable than trying to check every piece of their work.
RK: The six most expensive words in running a finance organization are “we’ve always done it this way.” Last year I wrote a piece about continuous accounting, it was a short essay on changing attitudes towards management finance organizations. Both accounting, FP&A, and treasury. The message to them was to use technology to automate as much as possible to cut down their own useless work and part of the process was having the continuous improvement mindset to offset the usual inertia we find any large organization, but which is particularly pervasive in finance and accounting organizations.
People who train in accounting do so because I think they’re happy with the idea of doing the same thing over and over and over the same way. That’s goodness and best practice in accounting. So you need to counter that with a mindset that says: let’s go through and examine what we’re doing and toss out the stuff that’s stupid, and create a mindset within the finance organization that says we’re going to stop doing stupid things and we’re going to figure out how to do thing better. That will lead to a review of everything that you’re doing, and that starts at the top. That starts with a CFO who’s fearless and wants to shake things up, which is not always easy to find. The way that you unleash a lot of the technology which you already have is by working on the people side of things, changing mindset, and that will free up time.
PH: We took an annual budget and then we went to rolling forecasts and made it monthly, but it’s still a batch process. That’s still a periodic batch process, and I think what we need to do is to really to make it continuous. A couple of things on that. Modern integration technology lets us stream data continuously into the application, so you can plumb data from your key business system, your operational system, your sales system, your ERP system – you can plumb that data with modern integration technology. The other thing we can do is we can automate it further with predictive analytics and machine learning, so we can actually automate this whole process. So if we can’t find the time to do it, rather than use a batch process, we need to use the latest technology to create continuous forecasting.
SP: But you still need to be careful. Because even if I do that with machine learning, the question is whether anything has changed significantly. If you’re going to do a rolling forecast, do not go to a monthly forecast. You want to look at what changes in the business. If I’m going to machine learning, one of the first things I’m going to check is what is the noise in the system. Even though I can see how my drivers are doing -and update my forecast based on drivers – in many cases what you’re going to find is a fluctuating forecast. That’s just noise. Within the wide range of statistical outcomes, we’re still within the standard deviation. What that tells you is that nothing has really changed other than there’s a minor fluctuation. There are things we know today that we are beginning to scratch the surface of what these systems can do. It gets down to understanding the algorithms and the predictive logic. If this happens, then this happens – once I’ve tested the algorithms and know what I can trust, then I have some sense of where we’re going. And if you have a sense of where you’re going, then all you really need to know is whether anything has changed. Is the standard deviation going upwards or downwards?
Ben: Finally, let’s give our analysts one final parting shot before we close today’s discussion. Steve, why don’t we start with you?
SP: Very simply, in finance, we have an opportunity to become more strategic by stopping doing dumb stuff. Get ourselves off the back of the boat and start looking forward. We’re putting in a lot of mechanics so that we can truly free up time and look at the physical things. Everything that happens in a business starts physically first. We can dollarize it later. If all we’re doing in accounting is waiting until we dollarize things, that’s all the important things that happen. You won’t be able to see it, what causes something to happen, what causes inform the predictive logic and work your way all the way back to the start. Typically I start with marketing and a lead and bring it through the system. We begin to understand the future that is already happening. That’s what we’re trying to get to; to use this technology to understand the future as its happening and understand what the trends are that tell us what could happen.
PH: Because you’re here, you’ve embraced software-as-a-service. What I like to think about is moving the rest of your systems to software-as-a-service. Some of the data that we collected shows that finance and ERP are really starting to move from an adoption level of about 6% a couple of years now to 19%. And 24% are planning to replace their finance and ERP systems within the next couple of years. So take your entire portfolio of business applications and leverage the benefits of software-as-a-service across your financial accounting system, across your ERP system, your purchasing system, your HR system, and so on.
RK: One of the easiest diagnostics for FP&A and finance organizations to figure out in terms of what they probably want to address in the way of technology processes and people issues is the close. If you want to figure out what’s out what’s wrong with your finance organization, one of the easiest way to do it is – how long does it take you to close your books? If it’s more than a week, there’s an issue. And if it’s not, well, maybe there’s still some improvements to be made. I’ve used that to identify the drivers. If it takes you more than a week, statistically our numbers show that’s half of you in this room. If it’s taking you more than a week, all of the things that are driving that result are probably driving a lot of the finance and accounting processes. Use that as a test. Then when you identify the sources, fix them and then you will probably be able to improve the performance of the finance organization, and the FP&A organization meaningfully.
Ben: Thanks to our panel. Let’s give them one final round of applause.