How To Use Data To Outsmart Your Competitors
The pressure's on to use data to outsmart your competitors. Here are six ways companies can use data to imagine and even re-imagine what's possible.
"Business as usual" can be a risky business practice, especially when there's cultural resistance to change. While some companies are embracing agile practices, there are a number of data-related barriers that keep companies from reaching their potential, most of which have to do with people, processes, and technology.
Industry disruptors make a point of using data in ways that other companies have not mastered, or may have not even considered. According to McKinsey & Company, most established businesses have not yet achieved impact at scale.
The clock is ticking, though, because the ability to leverage data is becoming less optional with time. For one thing, the competitive pressures have become obvious in virtually every industry. In addition, having a coherent data strategy is becoming a requirement on Wall Street. According to a forthcoming survey by Institutional Investor and KPMG, 24% of analysts have already revisited their opinion of a company or revalued it based on its data and analytics strategy. By 2018, that number will nearly double to 45%.
"When you think about the 6,000 publicly listed companies in North America, that's about $30 trillion of market cap," said Marshall Toplansky, managing director of KPMG Advisory Services, in an interview. "It seems it's going to be a very high-stakes game. Change is going to be forced on [public companies] if they don’t embrace it themselves."
Corporate culture is a major obstacle, however. Most companies are so vested in perpetuating their current business model that they want everything having to do with data and analytics to tie back to that business model, Toplansky said. They're also clinging to age-old decision-making processes.
"[For] most companies, when you think about how decisions are made, what comes to mind is a bunch of people who are smart sitting around a conference room table, diving into new pieces of information that they might not have had before in order to create new insights that would allow them to make a decision that might be different from a decision they made before," Toplansky said. "Data analytics isn't about insights. It's about speed."
There is also a misconception that more of the same data types will enable companies to magically realize significant business improvements -- even though it's outside data that's driving the data analytics revolution.
"The view is limited by the inputs and the data that [decision makers have] been using to model it in the past," Toplansky said. "It's being able to pull in unstructured, semi-structured, and structured data from outside data sources, put them into models, and see how they correlate with performance, that opens up companies to new ways of thinking."
Upstart disruptors, and even some very mature organizations, are using data to inform their business models, and in some cases that requires radical shifts in mindsets and corporate cultures. Other companies may ignore or disregard signals that are inconsistent with the current business model or the current way the company operates. Leveraging data requires adjustments at several levels. However, many companies are still struggling with data-quality issues and information silos that hinder their ability to imagine what can be done with data. Here, we reveal six strategies for coming up with creative ways to use big data.
It's hard to imagine the potential of data when the scope of data assets is not understood. In some organizations, important information is still locked away in silos, or else only a few of the people who should know what data is available actually do know. Alternatively, the scope of data assets may be known, but because some or all of them are not available in a timely manner there are fewer ways to leverage those assets.
For example, a lot of companies use online surveys to learn more about their customers and prospects. Online services such as Survey Monkey allow users to create, manage, and analyze surveys. The survey data gathered by a particular company provides valuable introspective information about that company, but there is also aggregate data across companies that can be used to gauge competitive performance. Survey Monkey collects about 3 million survey responses per day, and it has been collecting data for more than a decade.
"In the past five years we've been thinking about how to use the data we collected," said software engineering manager David Wong in an interview. "We had to rethink how we would use that data, [but] the engineering techniques at the time weren't scalable enough [and] you'd have a latency of six months to about a year between the collection, aggregation, and generating of the benchmark product and delivery. If you're looking for immediately contextual information to make a better decision, that's slow. From our perspective, it was almost a big step to say, 'Why can't we make this available in real time?' By asking those sorts of questions, we started re-imagining what the status quo might be."
In the future, Survey Monkey will use natural language processing (NLP) to analyze verbatim textual responses, which will enable additional capabilities. Wong said such initiatives require a combination of people -- those with data science and engineering expertise who can understand what's possible to do with data, and product-minded folks who can imagine how future surveys might differ from today's surveys.
Data has traditionally been IT's responsibility. And historically, many companies have limited the use of business intelligence (BI) and big data analytics to certain groups of people. Limiting the scope of data access limits the potential of how the data might be used. Moreover, the failure to involve certain parties -- such as the business users who might use an application, solution, or dashboard -- can cause organizations to invest in solutions or to embark on initiatives that are not as effective as they could be. Of course, security and privacy are valid concerns that need to be addressed and managed. However, it is easy to use security and privacy as an excuse to cling to old habits.
"[E]nterprises need to loosen up a bit," said Matthias Steinbauer, a computer scientist and researcher at Johannes Kepler University, in an interview. "Most of the startups I talk to have a policy [that allows] almost everybody to access data lakes. However, one needs to consider that many startups focus on IT, software, or apps, which means that a large portion of their staff is [comprised of] computer scientists who have a good understanding of technology. In other enterprises, it will be far more difficult for domain users to understand big data."
Usually, when people think about sharing data across the enterprise, they think in terms of dashboards and data visualizations. Shutterstock, a stock photography, video, and music marketplace, started an internal email campaign called "The Daily Dose," which tells employees about the data that's available and what's being done with it.
"We wanted the company to think in terms of data -- not just the data analysts and data scientists, we wanted developers, user experience, the sales team to be exposed to the data and data insights," said Sepehr Sarmadi, a data scientist at Shutterstock, in an interview. "We have intellectually challenging questions, and we wanted everyone to think about those problems. You never know who in the company is going to come up with the most clever idea. We didn't want to create wiki pages no one reads."
To be effective, services such as Shutterstock need to serve up relevant content instantly. In the past, if a search yielded no results, the landing page said exactly that. As a result of The Daily Dose, the landing page now displays an invitation to explore categorized collections that are displayed beneath the message. Meanwhile, the marketing department identified content gaps in which a critical mass of images were not yet available for a particular location, for example. To help fill the gaps, Shutterstock launched a new program that features one contributor from a different country each week.
All companies have data, but not all of them are realizing the full potential of it. For example, mobile operators collect massive amounts of information about customer behavior that can be used to improve internal performance -- whether that's optimizing the infrastructure, improving up-selling and cross-selling, or enhancing customer service. They are also monetizing data and selling it to third parties who are using the data to improve the performance of sales and marketing campaigns.
KPMG helped one of its clients understand how real-time traffic data could be used to improve the competitive success of promotions.
"Whenever you look at promotions, you ask questions such as, 'Did I get lift? More sales than the last promotion? What's my cost per incremental sale?' But what you'd love to know is whether the promotion actually stole market share from a competitor," said Marshall Toplansky, managing director of KPMG Advisory Services. "And you've never had the data to do that. Now imagine you're a brick-and-mortar retailer and you're able to see whether people are going to your competitor's store or your store -- and you can see that in real time and correlate it against when the promotions were sent out. Suddenly you have the ability to have a set of metrics about marketing accountability you haven't had in the past."
Everyone wants unicorns and data scientists, but not everyone can find or afford them. Even companies that have access to such multi-skilled talent limit potential accomplishments by expecting too much from a single individual or team.
"It's a common misconception that data projects are solely driven by data scientists," said Pauline Brown, director of marketing at Dataiku, a data science software provider, in an interview. "In reality, on an enterprise level, getting from raw data to actually deploying data-driven solutions [requires] many different skill sets. From the IT team, to the data wardens, to the data scientists, to the business decision-makers, and [to] all the executives, everyone contributes to the project. Unfortunately, getting all these people to work together is difficult." However, it is also necessary.
If you want different results, change the approach. Quite often, companies are stuck in traditional mindsets that discourage exploratory thinking outside of R&D, the data science team, or some other microcosm within the organization. While it's true that the vastness of big data can result in unbridled exploration that has little or no relevance to business objectives (and therefore little or no ROI), when experimentation, discovery, and brainstorming are aligned with a strategic goal, it is easier to focus on that which enables positive change.
Deriving value from big data can be like organizing a hoarder's garage: There's a lot of noise, and not always a clear signal. Starting with a business goal helps keep big data and BI initiatives on track, from the earliest conceptual stages to execution and refinement.
"The projects which seem destined to fail are usually too broad and open-ended, where the stakeholders start with a goal of 'analyzing data' as more of a research project and less of a goal-oriented effort," said Derek Gabbard, president of big data analytics solution provider FourV Systems. "It's easy to discover correlations in data when you don't know what you are looking for, and it is dangerous to rely on these discoveries without sufficient control. If you start with an end in mind, you can engage on a path that allows new discovery while validating the causal relationships and eliminating false positives. Companies that have a large body of knowledge already codified in a decision-support system can easily use this to validate big data analysis results and then slowly add unknown data relationships and causal effects to the process."
Deriving value from big data can be like organizing a hoarder's garage: There's a lot of noise, and not always a clear signal. Starting with a business goal helps keep big data and BI initiatives on track, from the earliest conceptual stages to execution and refinement.
"The projects which seem destined to fail are usually too broad and open-ended, where the stakeholders start with a goal of 'analyzing data' as more of a research project and less of a goal-oriented effort," said Derek Gabbard, president of big data analytics solution provider FourV Systems. "It's easy to discover correlations in data when you don't know what you are looking for, and it is dangerous to rely on these discoveries without sufficient control. If you start with an end in mind, you can engage on a path that allows new discovery while validating the causal relationships and eliminating false positives. Companies that have a large body of knowledge already codified in a decision-support system can easily use this to validate big data analysis results and then slowly add unknown data relationships and causal effects to the process."
-
About the Author(s)
You May Also Like