WEBVTT 1 00:00:00.000 --> 00:00:09.611 So what's the role of technology in improving forecast performance? Around about 2 00:00:09.612 --> 00:00:19.224 80% of organizations say they are using spreadsheets or they are using spreadsheets coupled 3 00:00:19.225 --> 00:00:29.837 with pivot tables and a little bit of BI to produce their plans, budgets and forecasts. 4 00:00:29.838 --> 00:00:38.449 So not too many surprises there, we all know that spreadsheets are used to a considerable extent, 5 00:00:38.450 --> 00:00:48.062 either in a very basic way or with pivot tables and BI, and we see a bit of an improvement in performance with 6 00:00:48.063 --> 00:00:57.675 BI and the use of pivot tables as well and we call those users 'advanced users' whereas people 7 00:00:57.676 --> 00:01:07.288 just using spreadsheets we call 'basic technology users'. But we also define the people who are using cutting-edge technology 8 00:01:07.289 --> 00:01:16.900 by which we mean people using graphing techniques for analysis, data visualization 9 00:01:16.901 --> 00:01:27.513 and charting to be able to get insights from the data, and then finally we looked at 10 00:01:27.514 --> 00:01:36.126 the experimental technologies, these are the people who are using artificial intelligence 11 00:01:36.127 --> 00:01:45.739 and machine learning to support their planning, budgeting and forecasting process and we only found about 12 00:01:45.740 --> 00:01:55.351 14% of organizations in this category, they tend to be the organizations with more than 13 00:01:55.352 --> 00:02:04.964 10.0000 employees who perhaps have the time and resources and their deep pockets that allow them 14 00:02:04.965 --> 00:02:14.577 to experiment with these newer kinds of technologies. And what we are going to do in a moment is look at 15 00:02:14.578 --> 00:02:24.176 how these different approaches affect the performance of the PB&F process. 16 00:02:26.094 --> 00:02:35.789 Attitudes to artificial intelligence and machine learning vary quite considerably, I think it's fair to say that the jury 17 00:02:35.790 --> 00:02:45.485 is out as far as most Finance professionals are concerned. Above the line in this graph we can see that 18 00:02:45.486 --> 00:02:55.182 56% of CFOs have acknowledge that we're going to become much more dependent on machine learning but also 19 00:02:55.183 --> 00:03:04.878 agree that humans introduce too much bias into forecasting, so if you like, a little over a half are really saying: 20 00:03:04.879 --> 00:03:14.574 'Yes we can see a role from machine learning and artificial intelligence because what we're doing isn't good enough'. On the other hand 21 00:03:14.575 --> 00:03:24.271 below the line you see people running back from the problem, saying that humans will always play a greater role 22 00:03:24.272 --> 00:03:33.967 than machines in forecasting, the global uncertainty will always limit the usefulness of machine predictions 23 00:03:33.968 --> 00:03:44.664 and therefore only 29% of CFOs agree that we should place more reliance on machine generated forecasts 24 00:03:44.665 --> 00:03:53.360 and ultimately only a quarter of CFOs feel that machines will be better at predicting the future than humans. 25 00:03:53.361 --> 00:04:03.056 So, how do the different technologies affect the performance of the 26 00:04:03.057 --> 00:04:12.753 BP&F process? Let me remind you that here we're looking at performance from the perspective of the ability to re-forecast 27 00:04:12.754 --> 00:04:25.449 quickly, the ability to re-forecast accurately between plus and minus 5% of earnings 28 00:04:25.450 --> 00:04:32.146 and then finally the ability to forecast beyond the one-year time horizon. 29 00:04:32.147 --> 00:04:41.842 What we find is that experimental technologies on the far right hand side here, 30 00:04:41.843 --> 00:04:51.538 they actually give a little bit of an edge in terms of the speed of re-forecasting 31 00:04:51.539 --> 00:05:01.235 if you use these technologies I suppose it's not too surprising because of the high levels of automation here and the ability to change 32 00:05:01.236 --> 00:05:10.931 quickly different scenarios, then machine learning and artificial intelligence seem to have just a sliver of an advantage 33 00:05:10.932 --> 00:05:20.628 over the other technologies when it comes to the speed of turning around and re-forecast, 34 00:05:20.629 --> 00:05:30.324 but in the other critical areas, in other words the ability to re-forecast accurately or to look beyond 35 00:05:30.325 --> 00:05:40.020 the one-year time horizon, then it is cutting edge technologies that have the upper hand and if you remember 36 00:05:40.021 --> 00:05:49.717 cutting edge technologies were the data visualization techniques, the charting and graphing 37 00:05:49.718 --> 00:05:59.413 that allowed people to drive more insight out of the forecast and basically cutting edge technology 38 00:05:59.414 --> 00:06:09.088 trumps the basics spreadsheets and the slightly more advanced users spreadsheets with pivot tables and BI. 39 00:06:10.645 --> 00:06:18.301 So what is it that defines cutting edge organizations? Those organizations are more likely 40 00:06:18.302 --> 00:06:27.958 to have moved to the cloud and they're more likely to have used specialist BP&F systems. 41 00:06:27.959 --> 00:06:33.614 So I suppose the interesting question out of this research is: 42 00:06:34.816 --> 00:06:44.312 'what is the most influential aspect in terms of improving the performance 43 00:06:44.313 --> 00:06:53.809 of BP&F? Is it being a more insightful organization or is it using more cutting edge technology?' 44 00:06:53.810 --> 00:07:03.306 So that's what we did, we compared those 2 approaches and we know that both insightful organization 45 00:07:03.307 --> 00:07:12.803 have a huge impact on the performance of BP&F, but also using cutting edge 46 00:07:12.804 --> 00:07:25.300 technology has a large impact on insightfulness and the speed of BP&F process 47 00:07:25.301 --> 00:07:31.797 and its accuracy and the ability to look out further in the year, and what we find is 48 00:07:31.798 --> 00:07:41.294 that cutting edge technology whereas improves performance a lot doesn't improve it as much as 49 00:07:41.295 --> 00:07:50.792 being an insightful organization can improve it. In other words what we draw from this is technology alone 50 00:07:50.793 --> 00:08:00.289 isn't the answer, what is important is you're using non financial data, that you're using rolling forecast, 51 00:08:00.290 --> 00:08:09.786 that you're using specialist software and then if you like the icing on the cake is using the 52 00:08:09.787 --> 00:08:19.283 cutting edge technology, the data visualization and analytics and charting and graphing. 53 00:08:19.284 --> 00:08:28.780 So it's a combination, it's about using technology and using some other techniques as well. 54 00:08:28.781 --> 00:08:38.277 In other words insight is a balancing act, it's about using the right techniques like 55 00:08:38.278 --> 00:08:47.774 rolling forecast and non financial data as well as cutting edge technology. 56 00:08:47.775 --> 00:08:57.258 I hope you enjoyed this webinar and look forward to welcome you to other webinars from FSN in the future.