Don’t assume your analytics partners will still be effective as you embrace smart analytics

Chief data officers, chief analytics officers, and business function leaders are being pressured into getting ahead on new opportunities with “smart analytics.” Our research on 153 enterprise operations leaders (see Exhibit 1) already shows that 60% of firms are on a two-year trajectory to infuse machine learning (ML) across their business lines as a core technology.

Exhibit 1: The race is on: two-year timeframe for mass adoption
What is your estimated timeline for ML becoming a core technology
implemented in all relevant aspects of your business?

Source: HFS Research, “How to avoid your looming machine learning crisis”, 2018

Sample: n=153enterprise leaders

Increasingly, business and IT leaders are turning to their data and analytics partners to help them design and implement smart analytics solutions that take them toward a data-driven future. But not all analytics engagements are created alike. This POV outlines how today’s smart analytics clients are finding the most success with providers that have cultivated the ability to be nimble and collaborative and bring a diverse set of skill sets to the table to guide clients on new technology changes in the next two years.

Analytics clients speak out: It all comes down to smarter talent development and engagement style

We define smart analytics as solutions that sift through unstructured data and make meaningful use of new technologies, such as using machine learning to augment human decision making and improve recommendations over time. In short, smart analytics systems can sense, comprehend, adapt, and recommend.

The HFS Blueprint Report: Smart Analytics report featured a detailed analysis of the capabilities and vision of analytics services providers as they adapt to new enterprise demands around smart analytics. This POV deep-dives into the key demand-side perspectives that emerged from client interviews from our Blueprint research. The primary research is based on interviews with 36 analytics executives, senior manager level and above, with representation from multiple industries across the globe.

Summarizing qualitative comments from a wide variety of operations and IT client executives, we found that the top-of-mind service provider strengths are the availability of relevant talent and skill sets (31% cite it as the top provider strength). Clients seeking smart analytics solutions need a wide range of skill sets, and thus greatly value their service providers when they can fill these gaps. Enterprises seek multi-disciplined analytics talent:

• Data engineers with specific platform skills, such as MapReduce, Apache Pig, Apache Hive, and Apache Hadoop;
• Data scientists coming from a variety of fields including PhDs in statistics, math, and computer science;
• Data visualization experts with backgrounds in statistics, graphic design, and UX; and
• Data analysts with experience in different programming languages, increasingly Python.

Related, analytics teams that understand business context, possess industry or horizontal-specific domain knowledge, or are familiar with a client’s technology and data setup are also in the top three strengths. As the director of data management and infrastructure at a global bank recounts, “We would not be able to hire the number of people we need with these skills… [Service provider] is also good at staying aligning its organization’s goals with ours so that the team’s behavioral outcomes are consistent and benefit both parties.”

Conversely, clients see the lack of talent development as the biggest weakness for some service providers. These include not having a global talent presence, the right technology skill sets, or the proper sourcing or retention capabilities for high-value analytics talent. Simply put, if you’re a service provider that is really good or really bad at managing analytics talent, your clients will either love you or leave you for it respectively. Unless service providers are developing the right culture for attracting and retaining multi-disciplined analytics problem-solvers, they are playing in the analytics marketplace of yesteryear.

Exhibit 2: Talent expertise and engagement styles are the biggest drivers of successful analytics

Q: What are a particular strength and a particular weakness or challenge for your service provider?

Source: HFS Research, 2018, n=36 analytics services client executives

Engagement styles, delivery structures, and client-oriented culture also stand out as top reasons for choosing to work with certain service providers. Collectively, these reasons represent the “right attitude” in Exhibit 2. The key attributes of the “right attitude” for analytics service providers, as perceived by clients, are

• Pragmatism: As one client describes, “They understand something that some analytics vendors don’t. They know they need to know the data, so they don’t gloss over the realities of the heavy lifting.”
• Agility: An online retailer that works in a fast-paced environment particularly values this, citing that their provider’s key valued strength is, “Working agile, with the commitment to deliver results even if it requires them to change their tools and systems.”
• Flexibility: This attribute is particularly important when it comes to pricing and contracting. Funding for various analytics projects ebbs and flows and technologies change, as do strategic priorities and the level of executive sponsorship and attention. Clients deeply value service providers that can flexibly navigate these challenges and find creative ways of solving them.
• Collaboration: Collaboration reflects how a service provider develops a partnership-based learning culture with clients. One client states simply, “I like working with them; they are good collaborators.”

Enterprises increasingly value — and see a gap around — the use of intelligent automation across the analytics lifecycle

The Smart Analytics Blueprint client surveys directed analytics clients to score their current engagements on a 1 to 10 rating scale across various dimensions of execution and innovation. The individual scores provide insights into each service provider’s capabilities, guiding Blueprint grid placement. At an aggregate level, however, these scores give us market guidance for how the industry is advancing as a whole and what areas clients view as collective pain points. Two areas of relative weaknesses thus emerged, with service providers scoring low on:

• Ecosystem development: We define ecosystem development as how “your service provider is developing its ecosystem around data and analytics to solve your critical business problems.” Enterprise clients expect leading service providers to bring outside-in perspectives and partnerships and provide access to cutting-edge developments in smart analytics. Examples of ecosystem development that clients find most providers lagging on include participating or investing in consortia, collaborating with academia, partnering with niche startups to vet them for enterprise use, and connecting with data providers that could add value to existing projects. As the senior vice president of digital analytics at a large financial institution recounts, “We don’t really see ecosystem development…it is always useful. As part of my broader role, I am constantly evaluating startups, looking at how different technologies can help us. Anyone who can bring that as a value-add would be an absolute help as a partner.”
• Embedding intelligent automation: We define embedded intelligent automation as “the level to which your service provider provides opportunities for and succeeds in embedding intelligent automation capabilities into data, reporting, and analytics solutions.” For HFS, intelligent automation refers to robotic process automation, autonomics, cognitive platforms, and machine learning. We see strong potential in embedding these technologies across the data and analytics lifecycle. Clients in our research agreed, but they also pointed out automation capability as a current industry gap. Early examples include introducing automated data ingestion and classification and cataloging, robotics to auto-populate reports, and machine learning algorithms to augment traditional models.

Exhibit 3: Average client scores reveal low confidence in automation and ecosystem development

Source: HFS Research, 2018, n=36 analytics services client executives

The data explosion is going to continue to impact businesses, as disruptive companies keep sprouting up with digital business models built around smart analytics. The gaps that analytics clients perceive around developing broader ecosystems and having embedded automation relate to countering these visible threats across industries. The marketing director at a global consumer electronics brand outlines the future potential of these technology shifts, “[Service provider] is just getting started and will need to start thinking big. Automation will be big, and we certainly need more with changes in our market landscape. When you talk about connected homes or mobility, we need more automated solutions around data and analytics. Data is growing, and we can’t do manual work to track and fix those things anymore.”

In a similar vein, projecting two years out (Exhibit 4), analytics clients expect service providers to bring more outside-in innovation to keep them ahead on technological change, whether that is harnessing unstructured data, moving to cloud platforms, or using automated solutions. Above everything else, analytics executives need help with ML and AI initiatives and expect service providers to rapidly develop expertise across ML consulting and implementation.

Exhibit 4: ML and AI expertise trumps all other critical success factors

Q: What are the key success factors you would expect to see from service providers in 2020? 

Source: HFS Research, 2018, n=36 analytics services client executives

Recommendations to push your engagement toward smart analytics

Our recommendations for partnering for smart analytics include:

• Go beyond “better math” with intelligent automation technologies: Several analytics leaders in our research pointed to the planning and piloting of machine learning and neural network-based algorithms to augment existing data models. However, there is still a lack of wider acceptance on the use of intelligent automation technologies to improve the data, reporting, and analytics processes themselves. Bar a few pioneering examples, we see a strong opportunity for services buyers to introduce modern automation tools to speed up and improve the quality and accuracy of connected data and insights.
• Challenge your service provider on the assumption of “retraining” talent for smart analytics: Large service providers manage vast estates of legacy work for enterprise clients in the areas of data management, reporting, and traditional statistical analysis. While they are all investing in creating new capabilities with changes in technology platforms and languages, the emerging narrative has been around retraining talent to take on new tasks. HFS cautions clients against this strategy; the talent availability and skill sets for traditional ETL and BI work or statistical analysis are significantly different from new needs around data engineering and machine learning. A segment of a service provider’s workforce may be upskilled or retrained, but for the most part, this industry will need to create and infuse new talent to take on these roles. Work with the service providers that are realistic and pragmatic about these changes and truly committed to closing the talent gap for the future.
• Drive governance and coordination across data, analytics, and automation CoEs:HFS sees a resurgence of strategic investments and modernization of data and analytics CoEs across enterprises. Many enterprise clients in our research are in the transitionary phase of moving from regional hubs or limited data centralization efforts to creating a focused analytics COE to drive organization-wide data strategy. The operating model for delivering smart analytics will need to build on the existing structures already created in the past, augmented with better coordination across other Triple-A Trifecta technology initiatives.

Overall, our primary research reveals that clients value analytics service providers most for their quality of talent, their working styles, and the amount of domain knowledge and context they bring to solve business problems. These qualities can be found in some of the largest professional services firms as well as niche startups and analytics and ML boutiques. We thus see a range of service providers proliferating in the data and analytics market as clients gain confidence and find the right culture to collaborate on data challenges. The next step is where it gets tricky—getting strategic with unstructured data and expanding on ML and AI use cases to transition to smart analytics. As an enterprise analytics decision maker, look for partners that are starting to demonstrate these traits:

• Developing smarter, automated processes across the data and analytics lifecycle to speed up and scale the level of operational intelligence in the enterprise;
• Collaborating across a broad range of partners and thought leaders to innovate in new areas such as ML and AI and layering on industry domain knowledge in the form of unique models and solutions or data partnerships; and
• Investing in shaping and adding to the net-new industry talent pool of data scientists, analysts, and engineers and helping you further build out data science practices and discipline.

This is where we will truly start to see smart analytics engagements. As one client puts it, they are looking for their provider to have “the ability to cross-pollinate with best practices, benchmarking capabilities and ways of working, innovating and innovative uses of data technologies to solve business problems.”

The bottom line: As smart analytics and AI become a vital strategic necessity, you must have analytics partners that embrace a much more collaborative, ecosystem-driven approach. Re-evaluate your current partners on the new needs around unstructured data, ML, and automation towards running truly “smart” analytics.