All companies need to undertake research and development (R&D), if only to ensure that their existing product lines remain competitive. Such day-to-day innovation is often termed ‘incremental’, and is contrasted with the kind of ‘disruptive’ innovation that leads to entirely new types of product (there’s a parallel here with the Red Queen Hypothesis in evolutionary biology, which states that species must evolve continually in a competitive environment simply in order to avoid extinction).

Companies’ business models vary, and so it’s no surprise that their approaches to innovation vary too. In its Global Innovation 1000 reports, for example, Strategy& (the consulting arm of PwC) identifies three broad classes of innovation strategy: Need Seekers, Market Readers and Technology Drivers.

Need Seekers (Apple, for example) “make a point of using superior insights about customers to generate new ideas”, the goal being “to find the unstated customer needs of the future and be the first to address them.” Market Readers (Samsung, for example) “focus largely on creating products through incremental innovations to products already proven in the market. They use a variety of means to generate ideas; most involve closely monitoring their markets, customers, and competitors.” Meanwhile, Technology Drivers (Google, for example) “depend heavily on their internal technological capabilities to develop new products and services. They leverage their R&D investments to drive both breakthrough innovation and incremental change.”

ZDNet recently published a special report on Innovation: How to be a World-Changer, and in one of the articles (Innovation through the looking glass) I looked at patterns in absolute and relative R&D spending, and patent awards, among a small number of leading technology companies (Google, Apple and Samsung). I’ve now had a chance to add several more companies, and the new data reveal further interesting innovation-related trends.

The graph below plots the number of US patents awarded to seven top tech companies in recent years against the amount they spent on R&D, building in three years between spend and award as an estimate of the average time required to do research, apply for and be awarded a patent (the latest figures are therefore for 2011 on this chart).

There is clearly a positive correlation between R&D spend and the number of patents awarded three years later, both between and within (most) companies. However, there’s also a great deal of variation: across all seven companies (45 data points), R&D spend accounts for just 23.4 percent of the variation in patent awards. Companies’ R&D budgets vary widely, ranging from Apple at the low end ($0.71-$2.43 billion/year) to Microsoft at the high end ($6.10-$9.04 billion/year). There also appear to be two distinct clusters among the higher R&D spenders, with IBM and Samsung significantly above the overall trend line and Microsoft and Intel significantly below it.

Before we discuss these patterns, let’s look at two time-series graphs, for absolute and relative R&D spend for the seven technology companies over the last 10 years:

Most of the companies’ absolute R&D budgets (above) show an increasing trend, in line with rising overall revenues — IBM and HP’s graphs are flat because their revenues have been flat over this period too (average annual growth rates of 0.36% and 3.1% respectively). As noted above, there are big differences in the amounts involved: in 2014, for example, Samsung spent $13.4 billion on R&D, compared to HP’s $3.4 billion.

When we look at R&D spend as a percentage of a company’s revenue (above), two things stand out. First, the companies form two clusters, with Intel, Microsoft and Google spending much more in relative terms than Samsung, IBM, Apple and HP. The second notable feature is that R&D spending at Google and Intel has outpaced revenue, these companies’ relative R&D budgets rising from 9.8% to 14.9% and 13.2% to 20.7% respectively between 2005 and 2014.

What the statistics (may) mean

Here’s a stab at what may be going on. Generally speaking, the larger a company’s product portfolio, the more it’s likely to spend on R&D aimed at ‘incremental’ innovation (to keep existing product lines competitive). That’s why Samsung (which makes TVs, fridges and washing machines, printers, health and medical equipment, mobile devices, networking and telecoms equipment, processors and memory chips, and LEDs) greatly outspends Apple (which basically makes iPhones, iPads and Macs).

Overlain on that background is the company’s innovation strategy — and in particular, whether it can be classed as a ‘Technology Driver’, as this model will generally require a bigger slice of the revenue pie than either a Need Seeker or a Market Reader. Intel, Microsoft and Google clearly fall into this class, with relative R&D spends over three times those of the other companies examined. This may also explain the apparently low number of patents these companies generate per R&D dollar compared to the general trend (see the scatter graph, above): more fundamental research aimed at ‘disruptive’ innovation will generally cost more, and may take longer to reap dividends in terms of patent awards.

One company stands out as a special case: IBM. Despite moderate absolute (2005-2014 average = $6.1 billion) and relative (2005-2014 average = 6.2% of revenue) levels of R&D spending, Big Blue consistently tops the annual list of US patent recipients — which is why it stands out as an outlier in the scatter graph presented earlier. The explanation for this undoubtedly lies in IBM’s longstanding commitment to research and development, and a ‘patent everything’ approach to intellectual property.

If you’re interested in this area, there is an extensive academic literature on the relationship between R&D investment, patents and innovation. There are many other aspects that could be explored, including the quality of patents in addition to their quantity, and the uses that companies make of their IP portfolios.