**5.1 Investment in digital innovation and growth of revenue**

The importance of digital was also clear in firm innovation investment decisions. Firms in our sample also reported their investment in innovation was split: 37% went to digital innovations, 24% to non-digital, and 39% to hybrid technologies that combined them. 99% of companies said that they had increased their investment in digital over the past five years, with 39% saying that they were investing 'a little bit more', 52% saying 'more', and 8% saying 'a lot more'. These increases in digital investment are highly correlated with increases in firm revenue. Inspired by Arora et al. [20], we asked about the share of firm revenue that was being generated from new or significantly improved goods or services from the past two years. We find a strong correlation between these measures. **Table 6** shows that firms that invest "much more" in digital are 32 percentage points more likely to have more than 50% of their revenue coming from new/improved offerings, and 49 percentage points more likely to have more than 25% of their revenue coming from those offerings.


#### **Table 6.**

*Share of revenue from new or improved goods or services.*

It is important to note that the causal direction of this relationship is ambiguous. For example, investment in digital could generate new revenue, plans for new revenue (e.g., a marketing push) could call for more digital investment, or other plans (e.g., new product launches) could generate more revenue and necessitate digital investment.

#### **5.2 External innovation and firm capabilities**

King et al. [46] highlight the importance of accessing capabilities outside the firm for technological innovation, since "complete sets of technology- and non-technologybased resources that facilitate progress through all stages or the technological innovation process may be uncommon among firms." We observe this directly in our data, where only 8.5% said that their firm's internal resources and capabilities were a better match ("superior") for their innovation project than were their competitors. Another 40% said that their capabilities were equal to leaders in the field ("strong"), and 50.5% of the firms reported that their firm's internal resources and capabilities were only on par with many others ("reasonable"). Faced with these capability shortfalls, firms often look externally to find expertise for these innovation projects, as shown in **Figure 3**.

Consistent with a view that the primary purpose of using external innovation sources is to get access to skills that the firm lacks, we find that those firms who

**Figure 3.** *Variation in innovation sourcing by firm capabilities (n = 600 projects).*

rated their internal capabilities "reasonable" were only 34% likely to source that innovation internally, whereas when their capabilities were equal to leaders in the field ("strong") this number rose by 40 percentage points (statistically significant at p < 0.01) and when firm capabilities were superior to all others it rose an additional 22 percentage points (statistically significant at p < 0.01). Thus, for areas with the best match between capabilities and the desired innovation, 96% of projects were sourced *internally*, but for those with the worst match, 66% of projects were sourced *externally*.

Interestingly, much of the movement to new, fast-growing external innovation sources that we documented earlier seems to be driven by the desire to innovate in digital technologies while facing skill shortfalls in these areas. In particular, all of the most-successful projects that these fast-growing sources produced for companies were digital innovations. These sources also worked on digital projects where firms had particularly weak capabilities, as **Table 7** shows. In **Table 7**, we discretize capabilities into as good, or better, than industry leaders ("superior" or "strong" coded as 1) or not (coded as 0). That is, capability is a dummy variable for whether the firm's capabilities for an innovation project are cutting-edge or not. As **Table 7** shows, on average firms had cutting-edge capabilities for 68% of the projects that they did internally, but that number dropped 38 percentage points for innovations sourced from traditional external sources (suppliers, customers, and competitors) and 47 ppt for innovations sourced from new external sources (universities, third-party, startups, and crowd). As columns (2) and (3) show, these internal capability shortfalls are particularly pronounced in the digital technologies outsourced to new external sources, where the difference of 11 ppts is weakly statistically significant (p-value = 0.08), whereas for non-digital projects the 5 ppt difference is not statistically significant. That is, firms have particularly weak capabilities for the digital projects that they outsource to universities, third-party, startups, and crowd.


#### **Table 7.**

*Share of projects where firms' capabilities were as good or better than industry leaders.*
