Big Data: Getting a Better Read on Performance

February 2016

BIG DATA: GETTING A BETTER READ ON PERFORMANCE

The benefits match those of earlier technology cycles, but companies must scale up their data-analytics skills to reap the gains.

by Jacques Bughin

Over the past several years, many com- data analytics. When we evaluated its panies have avidly pursued the promised profitability and value-added productivity benefits of big data and advanced benefits, we found that they appear to analytics. In a recent McKinsey survey of be substantial—similar, in fact, to those executives in this field, nearly all of experienced during earlier periods them said that their organizations had of intense IT investment. Our results made significant investments, from data indicated that to produce these significant warehouses to analytics programs.1 returns, companies need to invest But practitioners have raised questions substantially in data-analytics talent and about the magnitude and timing of the in big data IT capabilities.4 returns on such investments. In 2014, for example, we conducted a poll of senior Yet we also found that while data-analytics executives and found that they had seen investments significantly increased value- only modest revenue and cost improve- added or operating profits, the simple ments from them in the previous year.2 revenue impact for consumer companies was considerably lower. This finding, Our latest research investigated the mirrored among B2B companies returns on big data investments for a on the cost side, appears to confirm the random sample of 714 companies intuition of executives struggling to around the world, encompassing a mix uncover simple performance correlations. of industries and company sizes typical The time frame of the analysis also of most advanced economies.3 Our seems to be important, since broader findings paint a more nuanced picture of performance improvements from large-scale investments in data-analytics business domains—operations, customer- talent often don’t appear right away. facing functions, and strategic and business intelligence. These were our Analyzing data analytics key findings: The research avoided overweighting technology companies, since many deni- Big data’s returns resemble those of zens of the C-suite say that “we know earlier IT-investment cycles. that digital natives capture big returns, History tells us that it takes time for but does their experience apply to new technologies to gather force and those of us who live in a hard-wired diffuse throughout an economy, universe of factories and distribution ultimately producing tangible benefits channels?” Operating profit was used to for companies.6 Big data analytics— measure returns, since it captures the the most recent major technology wave— impact of big data both through value- appears to be following that pattern. added productivity and pricing power The average initial increase in profits from (often resulting from better customer big data investments was 6 percent targeting). The data also allowed us to for the companies we studied. That understand other aspects of the returns increased to 9 percent for investments on these investments—for example, spanning five years, since the companies the advantages of being the first data- that made them presumably benefited analytics mover in a given market.5 from the greater diffusion of data analytics over that period.7 Looked at from another We took care to measure the returns vantage point, big data investments from technologies specifically linked to amounted to 0.6 percent of corporate big data and therefore considered revenues and returned a multiple of only analytics investments tied to data 1.4 times that level of investment, increas- architecture (such as servers and ing to 2.0 times over five years. That’s not data-management tools) that can handle only in the range of the 1.1 to 1.9 multi- really big data. Looking beyond capital ples observed in the computer-investment spending, we assessed complementary cycle of the ’80s but also exceeds the investments in big data talent across multiples others have identified for R&D eight key roles, such as data scientists, and marketing expenditures.8 analysts, and architects. Finally, we examined whether improvements were Investments are profitable across radiating throughout organizations key business domains. or captured only in narrower functions or Companies, we found, benefit broadly individual businesses. from big data investments. With minor variations, spending on analytics to gain Gauging performance competitive intelligence on future Our research looked at the results of big market conditions, to target customers data spending across three major more successfully, and to optimize

2 operations and supply chains generated example, by making IT-architecture operating-profit increases in the 6 per- investments in isolation. That’s a cent range. Although companies struggle mistake: about 40 percent of the profit to roll out big data initiatives across the improvements we measured resulted whole organization, these results suggest from complementary and coordinated that efforts to democratize usage— investments both in IT and in big data getting analytics tools in the hands of talent. Organizational constraints can as many different kinds of frontline make such gains difficult to achieve, employees as possible—will yield broad of course, since companies often silo performance improvements. their investments. For instance, the IT or technology department is commonly Understanding investment patterns tasked with determining the level of big data investments needed, while business Three aspects of big data investments units and HR departments draft their own determine the magnitude of these perfor- spending plans for employee resources. mance improvements:

We find that when companies fully Investing early augments the benefits. coordinate their investments in IT capital Our research helped us identify how with those in skilled roles, performance significantly early investments in big data improves substantially. Here’s an example analytics can raise the pace at which of what happens when they don’t operating profits improve: first movers coordinate them: one company’s large accounted for about 25 percent of investment in database-management the increase in our sample. One possible software foundered when HR neglected explanation is that early adoption allows to hire the analysts needed to support companies to learn by trial and error the new data-driven business priorities. how best to design data-analytics tech- Experience also tells us that in the nology and integrate it into their work- most capable organizations, a chief data flows. This, in turn, could create valuable or analytics officer often coordinates capabilities that help companies differ- IT spending with efforts to acquire entiate themselves from competitors. If analytical talent across business units. the cycle continues as increasingly powerful data-analytics applications Investing in big data talent at scale come on stream, the importance of is a must. rapid experimentation and learning—and Skilled employees across the spectrum of leaders who feel comfortable with of data-analytics roles are in short supply, this approach—could rise. so aggressive actions to address this problem are critical. Our study found that Combining investments in IT and 15 percent of operating-profit increases skills is decisive. from big data analytics were linked to the Many companies still compartmentalize hiring of data and analytics experts. their data-analytics initiatives—for Best-practice companies rarely cherry-

3 pick one or two specialist profiles to 1 In mid-2015, McKinsey polled 20 industry-leading address isolated challenges. Instead, they analytics executives on their investments to date. The results, while not scientific, were instructive: 90 percent build departments at scale from the reported medium-to-high levels of data-analytics start. With a broad range of talent, these investment, 30 percent called their investments “very significant,” and 20 percent said data analytics was companies can use data analytics the single most important way to achieve competitive to address the current challenges of their advantage. 2 See David Court, “Getting big impact from big data,” functional areas while developing McKinsey Quarterly, January 2015, mckinsey.com. forward-facing applications to stay ahead 3 These investments include a full range of spending on of competitors. big data software, analytics, hardware, and data- analytics talent. We used company data to calculate operating profits and value added. 4 Data were for the year 2013. For the complete set of findings and methodology, see Jacques Bughin, “Big data, big bang?,” Journal of Big Data, January 2016, Combined, these three investment journalofbigdata.com. 5 characteristics account for about 80 per- For additional analysis of big data returns, see Russell Walker, From Big Data to Big Profits: Success with Data cent of the operating-profit increases and Analytics, New York: Oxford University Press, 2015. in our study. Staying on top of new 6 In 1987, Nobel Prize laureate Robert Solow, who studied productivity effects of adopting computers, developments, carefully balancing invest- famously remarked, “You can see the computer age ments in skills and technologies, and everywhere but in the productivity statistics.” 7 becoming a magnet for cutting-edge In the operating-profit measure we account for the tendency of the most productive companies also to be talent will be the paramount considera- early big data adopters. tions for leaders keen to turn their 8 See Sunil Mithas et al., “The impact of IT investment on profits,” Sloan Management Review, March modest data-analytics gains into broader 20, 2012, sloanreview.mit.edu; and Sunil Mithas and more substantial ones. et al., “Information technology and firm profitability: Mechanisms and empirical evidence,” MIS Quarterly, 2012, Volume 36, Number 1, pp. 205–24, misq.org.

Jacques Bughin is a director in McKinsey’s Brussels office.