Does Innovation Capability Improvement Make Sense?
Editor | On 07, Jun 2018
Darrell Mann
Here’s the kind of innovation data guaranteed to make me smile:
Figure 1: Data And Crackpot Rigour (Reference 1)
Not so much because of the original intent – which was, I think, to establish possible correlation between R&D spend and business performance – but because of the decision to deploy mathematics to try and make sense of the data. Rather than trying to think about what’s going on.
The amusing part of the Figure, in other words, is the surreal hypothesis that sales growth equals 0.0116 times R&D sales ratio + 0.0832. While the equation might come from a mathematically valid (least squares) calculation of the results from the1000 companies included in the analysis, I can’t imagine there’s an accountant on the planet that is going to make any kind of meaningful use of the data. For a start the CTO would laugh at them.
If anything, what the graph is really saying is that there is effectively zero correlation between sales growth and sales growth. The mathematics is meaningless because it’s not possible to simply plot published company R&D sales ratio and growth figures and expect them to be meaningful. Sure, such data is readily available – companies are legally obliged to publish the relevant data in their annual accounts – but just because something is easy to obtain, doesn’t mean that it should make any kind of sense to correlate it.
There are several fundamental flaws in the logic shown in the graph. Here are three:
1)   R&D spend this year doesn’t translate to Sales Growth this year. There is an inevitable lag between the two things. This lag varies according to the ‘pulse rate’ of the industry. If the company is in a high pulse rate industry – say consumer electronics or telecoms – then the lag between R&D spend and Sales Growth might be quite low (less than a year right now). If the company is in a slow pulse rate industry, on the other hand – such as aerospace, marine or mining – the lag might be measurable in decades. Unless this industry-specific pulse rate is brought to bear in the calculation, the result will be meaningless.
2)   Even more difficult to isolate in the Sales Growth axis of the Figure 1 graph is how much of the growth can directly be attributed to R&D and the creation of new products and services. Versus, for example, the growth that is merely attributable to the way in which the economy is changing (the ‘rising tide lifts all ships’ effect) in general and how far up its evolutionary s-curve the specific products and services of the company in specific detail. These kinds of things are not published in company annual accountants. Mainly because the accountants have no idea how to compensate for these effects. Most, in my experience, think s-curves have something to do with driving their Porsche’s around dangerous bends in the Swiss Alps.
Some researchers have tried to solve this problem with another kind of crackpot rigour: Likert scale questionnaires (‘rank the level of aesthetic design capability in your organisation from 1 to 5’) sent to the wrong people (designers!) followed by an intricate but utterly meaningless mathematical analysis of the summed questionnaire results to produce pictures like this:
Figure 2: More Crackpot Rigour – ‘D.Index’ (Reference 2)
The only sensible way to try and isolate the ‘how much of the growth is attributable to the R&D spent in the organisation is to either get down-and-dirty with the C-Suite leaders of the organisation, or to have an objective means of plotting the maturity of an industry… such as Evolution Potential!
3)   The third logic flaw in the Figure 1 graph is an assumption that the innovation capability of each of the 1000 companies in the analysis is equal. Some companies are operated with a ‘pioneer’ business model. Others are happier operating as ‘fast-followers’. The former might benefit from greater technical innovation capability; the latter is likely to win by possessing greater business innovation capability… which raises an even bigger question of how do you calculate the amount of money invested in business innovation? That sort of information is far less likely to appear in the annual accounts. In no small part because most organisations have no idea what ‘business innovation’ is still.
The real research question behind the crackpot rigour of Figure 1 is not about what the average of the 1000 data points (the ‘average’ is, as ever, almost always meaningless in the innovation context) tells us, but what, if any, are the differences between the data points at the top of the graph versus those at the bottom? After, of course, the above three problems (and probably others) have been duly accounted for.
To be honest, it’s a question we’re still trying to plough through the details of. When you elect to do something ‘bottom-up’ instead of ‘top-down’, the amount of work you have to do increases exponentially. That’s why accountants prefer top-down. That’s why the results they publish are almost invariably meaningless at best and positively toxic at worst. Fortunately, these days the contradiction between top-down-or-bottom-up gets resolved by tools like PanSensic that allow us to scrape through lots of data from a bottom-up algorithm design perspective. The big analysis is still ongoing. But one thing has already emerged as extremely significant in terms of understanding the relationship between R&D spend and Sales Growth.
That thing is Innovation Capability Maturity. We got a first hint of the link from this picture:
Figure 3: EBIT-versus-Sales CAGR-versus ICMM Level
And here’s what the Sales Growth story looks like when we plot ICMM Level against pulse-rate and rising-tide/s-curve-maturity-corrected Sales Growth Figures (Figure 4).
Figure 4: Sales Growth Versus ICMM Level
(* Level 0 companies = start-ups. 75% of all innovation comes from Level 0 companies, but the success/failure curve is heavily skewed: 1 scaled success compensates for 9 failures hence the ‘average’ (mean) 6% sales growth figure needs to be taken with a fair sized pinch of salt – the good news is that innovation is overall better than a ‘zero-sum’ game, but the 6% average came from a small number of ‘rockets’ compensating for a much higher number of lemons.)
ICMM Level 1 companies, on average, will lose 6.5% sales each year. Not shown on the graph, but consistent with the original Reference 1 data, there is no correlation between sales growth (or ‘loss’ in this case) and the actual amount of R&D spend. ICMM Level 1 companies will, on average, shrink by 6.5% each year whether they spend nothing on R&D or whether they spend all of their income on it.
The lack of correlation between Sales Growth and R&D spend continues when we look at ICMM Level 2 organisations. Again, if you are ICMM Level 2, it doesn’t matter how much you spend on R&D, at least in terms of business performance. The good news, on the other hand is that your average Sales Growth will be around 4%.
This 4% figure rises to a whopping 22% for the ICMM Level 3 companies. And if you’re one of the relatively few Level 4 organisations on the planet, your average sales growth will be well over double that of the ICMM Level 2 companies, at a staggering 58%.
We’re still trying to unravel the Sales-Growth/R&D correlation for these two ICMM Level companies. We think there’s a positive correlation – i.e. greater R&D spend gives greater Sales Growth – but it doesn’t look like a linear relationship.
We’ll no doubt be publishing more on this story as the ongoing research analysis digs deeper. In the meantime, we thought it was a worthwhile exercise to publish the Figure 4 finding now in order to – hopefully – solicit some support from readers to either help verify (or not!) the finding inside their own organisations. Or (perhaps even better) to challenge the logic of what we’re trying to do.
The reason both of those things are important is that, if we’re right, the decision of a company to invest in Innovation Capability building suddenly becomes something of a no-brainer: if I can expect to increase my sales by over 10% by climbing from ICMM Level 1 to Level 2, that increase pays for an awful lot of training, innovation tools and methods. The only question then becomes one of time to payback… which, sadly for some, depends a lot on the prevailing pulse rate of their industry.