WEBVTT

1
00:00:00.000 --> 00:00:08.354
One of the main drivers of forecast performance and accuracy is non financial data.

2
00:00:08.355 --> 00:00:16.710
Although it's an old piece of research, probably more than 10 years old now, 'In the dark' was a seminal piece of research

3
00:00:16.711 --> 00:00:25.066
into the role of non-financial data in measuring the performance of an organization.

4
00:00:25.067 --> 00:00:33.422
At that time 78% of CEOs said that financial indicators alone simply don't adequately

5
00:00:33.423 --> 00:00:41.778
capture the company strengths and weaknesses, and we've really taken that concept forward and said: 'Well, to what extent

6
00:00:41.779 --> 00:00:50.129
can non financial data help with the performance of budgeting, planning and forecasting?'.

7
00:00:51.768 --> 00:01:00.742
What we find is that organizations that do harness non-financial data, more than half of them are able to forecast earnings

8
00:01:00.743 --> 00:01:09.718
between plus or minus 0.5% vs just 29% of organizations that don't make considerable use

9
00:01:09.719 --> 00:01:18.693
of non financial data.  They're also 2 and a half times more likely to be able to re-forecast 

10
00:01:18.694 --> 00:01:29.000
within 24 hours, so they're more responsive, they're 2 and a half times more likely to be able to respond more quickly to market

11
00:01:29.001 --> 00:01:36.645
change but most importantly, they're more than twice as likely to be able to forecast

12
00:01:36.646 --> 00:01:45.620
beyond the 12 month horizon. So non financial data is absolutely critical to being able

13
00:01:45.621 --> 00:01:54.596
to look out further into the future with our forecast, but what we find in our research is only 11% of organizations

14
00:01:54.597 --> 00:02:03.565
are making more use of non financial data now than they were 3 years ago.

15
00:02:04.792 --> 00:02:13.879
I suppose a bright spot in the research is that 78% of CFOs agree that the key to forecasting more accurately

16
00:02:13.880 --> 00:02:22.968
lies in the greater use of non financial data. When we ask organizations about 

17
00:02:22.969 --> 00:02:32.056
their use of data as a whole, what we find is that 78% say that decision making has become 

18
00:02:32.057 --> 00:02:41.145
more data driven over the last 3 years, but 58% concede that valuable data is

19
00:02:41.146 --> 00:02:50.234
scattered across the organization, 48% say that other functions are not good at sharing data with the finance function,

20
00:02:50.235 --> 00:02:59.318
only 40% have managed to formally identify how new sources of data could give them a competitive edge.

21
00:03:00.795 --> 00:03:10.518
So, the underlying thrust of this is that organizations appear to be more data driven but they're only more data driven 

22
00:03:10.519 --> 00:03:20.243
about the data they know about, they're not actually expanding the view of data into new sources of data

23
00:03:20.244 --> 00:03:29.968
and in particular non financial data. If you go back 10 years then probably about 80% of the useful information,

24
00:03:29.969 --> 00:03:39.693
the staff that could give us a competitive edge came from the general ledger and 20% came from outside that data universe.

25
00:03:39.694 --> 00:03:51.000
Going forward to now it's probably the reverse, I mean 20% comes from the non data universe, in other words our general ledger,

26
00:03:51.001 --> 00:03:59.142
and probably 80% of what's really going to matter competitively lies outside of the general ledger.

27
00:03:59.143 --> 00:04:08.867
And what's the role of continuous planning? What we see in this year's research is what we call the hamster wheel

28
00:04:08.868 --> 00:04:18.592
effect. 71% of organizations this year are re-forecasting more than twice a year

29
00:04:18.593 --> 00:04:28.317
compared with only 56% in 2016, in other words we're speeding the hamster wheel faster

30
00:04:28.318 --> 00:04:38.041
and faster, but the question is: 'Are we discovering anything new?', all you seem to be doing is repeating

31
00:04:38.042 --> 00:04:47.766
what we've done before, except more frequently, is almost like a comfort factor in the face of uncertainty and there has been a huge amount

32
00:04:47.767 --> 00:04:57.491
of uncertainty in the last year, we're taking comfort from the fact that where re-forecasting more frequently we appear to be

33
00:04:57.492 --> 00:05:07.216
showing people that we're doing something about it, but we're actually not impacting on our insight or our accuracy in terms 

34
00:05:07.217 --> 00:05:16.941
of forecasting. So, doing what we did before, just re-forecasting more frequently, just has a marginal effect on forecast

35
00:05:16.942 --> 00:05:26.652
accuracy where 1.25 times more likely to be able to forecast earnings between plus or minus 5%,

36
00:05:28.119 --> 00:05:37.214
but if those organizations have used rolling forecasts actually I can point to

37
00:05:37.215 --> 00:05:46.310
a much larger improvement, 43% of organizations that say their BP&F process is insightful use

38
00:05:46.311 --> 00:05:55.406
rolling forecast compared to just 29% who say their process is not insightful,

39
00:05:55.407 --> 00:06:04.499
in other words rolling forecasts are absolutely crucial to improving the insightfulness of the forecasting process.

40
00:06:05.922 --> 00:06:14.539
What else can we do to improve the BP&F process? In the last few years there has been quite a C change 

41
00:06:14.540 --> 00:06:23.158
in the capabilities of technologies: we can now build much bigger, more granular models without necessarily

42
00:06:23.159 --> 00:06:31.775
impacting on overall performance of the data model itself, but does it really improved anything?

43
00:06:33.691 --> 00:06:43.177
60% of CFOs say they're building much bigger models than they were 3 years ago, and they've much more complex models than they were 

44
00:06:43.178 --> 00:06:52.664
3 years ago, and it actually is helping with forecast accuracy? Those that use bigger models, 

45
00:06:52.665 --> 00:07:02.152
40% to them say they're able to forecast revenue between plus or minus 5%, they are 2 and a half times more likely to be able to react quickly

46
00:07:02.153 --> 00:07:11.639
to market change and the quarter of them say they're able to forecast beyond the 12 month time horizon. So bigger models

47
00:07:11.640 --> 00:07:21.123
do seem to have a beneficial effect on forecasting accuracy and the time horizon.

48
00:07:22.335 --> 00:07:31.644
What about adding more stakeholders to share in those larger models? The truth is that engaging with more

49
00:07:31.645 --> 00:07:40.954
stakeholders seems to improve the receptiveness, the acceptability of the forecast in the organization,

50
00:07:40.955 --> 00:07:50.264
but it doesn't really do much to improve the accuracy of the speed or the time horizon,

51
00:07:50.265 --> 00:07:59.574
so we find that engaging with more stakeholders leads to 2 and a half times the process being perceived more

52
00:07:59.575 --> 00:08:08.884
positively, being able to react more to market change and more people trust the data. So if we compare adding more

53
00:08:08.885 --> 00:08:18.194
stakeholders or building bigger data models, it's clear that both have a benefit: adding more stakeholders to the process

54
00:08:18.195 --> 00:08:27.498
builds trust but bigger models lets you see out further with more accurate forecasts.