Challenges in the integration of advanced data analytics

End decision makers, budget managers, and executives have been told for years that "advanced analytics" provide better answers to almost all questions in everyday business. Despite the ease of access, few companies in the retail sector are taking full advantage of all the opportunities digitization offers in analytics. Primarily because of the current situation in our urban ecosystems, one inevitably wonders why this is so. The need for valid answers outweighs any effort to use and implement advanced analytics. 

Amazon, Walmart, and several leading retailers at the top of the data pyramid use historic, dynamic, and even live data to make crucial decisions based on ever-changing customer needs. At the same time, the competition is still using simple tools that can only show the past rather than forecast the future. 

The real-world consequences for the industry are already being felt: the top 25 performing retailers - most of whom are digital pioneers - were about 83% more profitable than laggards during the pandemic. They took home more than 90% of the sector's market capitalization gains (source: McKinsey). It is estimated that in food retail alone - an industry with shallow margins - intelligent analytics would increase revenues by up to 2% - a potential money-maker for a challenging and complex business.  

Of course, none of this is news. But even corporate executives who have been slow to adapt to the new realities need to be aware, beyond a certain level, that they are leaving much money on the street here. Despite understanding its benefits and being aware that their competitors have quickly benefited in recent years as startups or innovators develop increasingly advanced analytics solutions, most laggards don't seem likely to catch up with the leaders anytime soon.

‍6possible reasons we have observed since 2019: ‍

WHATALOCATION's clients include many users, from expansion managers to financial experts, real estate agents, and shopping center operators. They all have different levels of analytics maturity and use our analyses for the most diverse use cases. Our network also includes companies that decided not to work with WHATALOCATION, even though our analytics would have had an immediate business impact. These six factors played a significant role in why companies currently decide against advanced analytics:

‍Company
‍Many companies have the challenge of maintaining the balance between centralization and decentralization. Centralization for efficiency, scale, and consistency; and decentralization for more flexibility, adaptability to local conditions, and receptivity to creative ideas to solve problems. ‍

Processes
Companies do not have unlimited resources to achieve their goals. Some of our interviewees noted that their analysis projects often take too much time internally to prepare, so the impact time is often exceeded. In other words, it is already too late by the time the result comes out because the market has evolved much faster. What is needed here are better processes with clear responsibilities for the overall goal.‍

Individual solutions
In the past, many companies have acquired a hodgepodge of legacy systems. Sometimes, they use three or four different solutions, but they only serve one use case. Unfortunately, these companies cannot keep up with the exponential growth of available data. Discrepancies between the data's sophistication and the tools' sophistication are also common.‍

Mindset
People often question whether data is important at all. Even if the answer is yes, some contacts didn't know why. Some companies suffer from risk aversion and have no clear goals for an analytics project. Others developed an almost condescending view of advanced analytics, viewing the process as an art rather than a science. A change in mindset is needed here, especially in Germany. ‍

Data, data, and yes, data
‍Data - the biggest problem. Nearly every company that fails at advanced analytics faces poor data quality and data management. Those companies manage data in different places or think exclusively about data silos. Some companies don't even collect the data they need. "There may be data about our customers from the customer loyalty program," said one executive. "[We] do have a customer app. However, we don't care about the information about their purchases or our customer base's demographics."

Yes, most employees in flagging companies are highly dissatisfied with their current situation and hope they can make a difference. They want to store data in the cloud, better monitoring, and invest more capital in better technology to improve the customer experience sustainably. They want support for decision-making at the micro level, store by store. But that would often require a lot of internal cleaning up. 

Some had told us they would like more help with modeling and solution strategies. They also want help integrating non-traditional data, such as mobility and demographic data, weather data, in-store customer activity, clickstreams, social media activity, or online search trends. But beware: it's just not enough to get the data. It would help if you had the tools to turn that data into actionable knowledge.‍

People
In our discovery meetings, we observe that analytics is driven by people who don't understand the business. Mid-market companies face a critical shortage of people with the right skills to develop and use analytics tools. These companies need people who can fill functional gaps - people who can act as translators between analytics and the business. So far, we have encountered such a person only once in a fashion company. Thousands of data scientists are needed in the market, but only a few hundred enter each year. LinkedIn lists more than 8,000 business analytics job openings in Germany, nearly 43,000 across Europe, and more than 94,000 in the United States.

So how should companies get started?‍

First, take stock of where the company stands:

What essential decisions do you make, and how advanced is the analytics used to create them? Are you organized in a way that allows individual units to experiment and innovate using analytics while learning from those experiences? Is your culture ready to embrace an evidence-based approach to decision-making? Do we have a systems infrastructure to collect, store, organize, access, and process all the information needed for analytics? Do you have people with skills that can translate back and forth - from business problems to analytics problems and then from analytics results to business recommendations? 

Additionally, ask what processes can be improved through better analysis using existing data: How can you improve the data needed for research? How can you make data more forward-looking and advanced? Once this initial assessment is complete, the real journey begins. And that path includes strategic investment and organizational redesign. 

Strategic investment

Leading retail companies invested significantly in their data and analytics infrastructure years ago. They have strategically decided against legacy systems and favor the transformation to cloud-based systems. This move solves the critical challenge of updating legacy systems: integrating the new and the old. Employees are fed up with incompetent interfaces that can't connect with existing systems. Cloud-based systems bypass such challenges and are designed to scale. They grow with the necessities of Big Data and Co.

Just as crucial as centralized storage is data quality. One of the identified challenges is the isolation of existing data. Data silos make it extremely difficult for employees to incorporate enterprise-wide data into decision-making. Decisions that provoke impact in the future to involve crossing organizational and departmental boundaries (e.g., expansion and marketing or from finance to C-level). Unconditionally breaking down data silos so that marketing teams, for example, can incorporate operational factors such as expansion plans in lead times can only improve business performance.

Most importantly, we see a strategic investment in talent and a pipeline of such vital people. Work with universities that offer data science degree programs. Students at these institutions are always looking for real projects to work on. Kill two birds with one stone: students gain practical business knowledge and can practice communicating analytical solutions to department heads. Conversely, the company learns about the latest tools and a real-world preview of some talent for future hiring opportunities. Think about training programs. Customized in-house programs can teach the C-level some of the basics of analytics and provide business domain knowledge to people in purely analytical roles.

Organizational redesign

"Think big, start small, and scale fast" - retail market leaders have a contemporary mindset and corporate culture that celebrates experimentation. Like WHATALOCATION, such companies consider data their DNA or a vital element. Shaping a modern corporate culture is no easy task, but we believe that a good first step can be made by reformulating the company's values with analytics.

Managers can lead an internal campaign emphasizing that analytics should empower decision-makers, not replace them. Employees should be rewarded for understanding analytics tools' predictions and recommended actions rather than simply executing the recommendations and rewarding compliance. It must become the first goal that data-driven decision-making becomes a vital cornerstone of the corporate culture.

Successful companies have complemented this experimental culture with a successful organizational design. They used a hub-and-spoke structure in which some expertise is embedded in specific business processes, and some reside in the center of excellence for analytics. A center of excellence can provide a community for those working on analytics, facilitate oversight, encourage knowledge sharing, and pool resources. Of course, some members of those teams should also be involved in operations. In this way, the company avoids the risks when the center of excellence operates in isolation. In particular, the risk of teams working on more technically appealing problems than practically relevant ones.‍

CONCLUSION‍

Technological (r)evolutions usually occur in two overlapping phases: the acquisition of tools and the acquisition of the know-how to use the tools. However, these two phases, developing the know-how to use the new tools, can also often slow adoption. At the beginning of Alexander Graham Bell's career, there were few telephone operators. In that respect, the data analytics revolution is no different. What is different is the speed at which these new tools are being developed. In this age of data abundance, those who act first to capitalize on their insights will surely be at a massive operational advantage over their competitors.