Data is coming in from different channels, mobile devices, IoT devices, to the point where it is often overwhelming. With so much data being produced and accumulated every minute, brands can begin to drown in data. That said, not all data is valuable or actionable, so brands need to be able to discern which is which.
“Enterprises are not drowning in data because of its depths, they are drowning because they don’t know how to swim. Too often organizations approach their analytics blindly, sifting through mountains of data hoping to find gems,” explained George Kobakhidze, enterprise solutions at ZL Tech, an information governance platform provider.
A Strategy for Obtaining Good Data
The amount of data that is produced in one day by people across the globe is staggering. According to Raconteur’s A Day In Data, on each day of 2019, 400 million tweets and 294 billion emails were sent, 4 petabytes of data were created on Facebook, and by 2025, it’s estimated that 463 exabytes of data, (463,000,000,000,000,000,000 bytes), will be created each day.
For brands that need to sift through all the data that is created through their various channels, including social media, email, websites, apps, sales inquiries, customer service tickets, surveys, and brick-and-mortar storefronts, a data strategy is required to separate good data from useless data. An effective data strategy allows brands to collect and analyze data, obtain actionable insights from the data, and store the data for future use. Often, this is where the problem lies — the lack of an adequate strategy. “The issue lies in that information is frequently stored without a strategy for future use, leaving data in the dark — unused and unknown. With upwards of 80% of enterprise data being considered dark, it is no wonder that analysts feel like they are drowning as they are unable to access relevant information, rendering the entire data collection process moot,” said Kobakhidze.
It’s also important to recognize that data has an expiration date — that is, data that was collected before the COVID-19 pandemic, for instance, is largely not useful for making actionable insights today. George Corugedo, co-founder and chief technology officer at Redpoint Global, a CDP platform provider, told CMSWire that brands need to remember that the overall principle is that data spoils. “So if you want to take the greatest advantage of the enormous volumes of available data (which everyone should try to because it is valuable) then all processes need to be performed under a ticking clock, reminding them that the data they are working on is dying a little bit every second.”
According to Walter McAdams, VP of solutions of Rhode Island-based SQA Group, a software quality assurance solution provider, brands need to structure a foundation in which their data strategy is business driven. In this way, it can provide insights and strategic guidance to the brand. The data must be trustworthy, reliable, accessible, and digestible to the software with which it will be used. "When creating a data-centric culture, you need to establish fundamental pieces: a data governance structure, clear understanding of ownership, unified data definitions, a mechanism for ensuring data quality, and methods for reporting, accessing and querying data that puts actionable information in the hands of stakeholders,” McAdams explained.
“If your foundational data structure is solid, you can’t overuse data and it reduces the possibility that you will misuse your data,” he said. “Instead, you can shift to turning data into actionable insights. Data serves as our compass: indicating what to do and where to head next.”
McAdams said that above all, brands should keep two things in mind: keep their data definitions up to date, and recognize that data quality never ends —i t goes on forever. “To get good data is essentially to get good data definitions that are endorsed by the stakeholders,” he said. “An operational data quality program can improve data ingestion and processing methods by monitoring how successfully they result in ‘good’ data in its rest state.”
Related Article: Why CIOs Need a Data Readiness Strategy
Put Technology To Work for You
According to a report from Salesforce, the average marketing organization uses 14 data sources, a number that is expected to rise to 45 data sources by 2025. The sheer volume of incoming data is both staggering and daunting, which is why technology must be harnessed to take control of it. Sandeep Kharidhi, general manager of data and analytics platforms at fintech services company, Deluxe Corporation, said that technology plays a key role in harnessing and making data actionable, as well as helping with governance and access issues. “While traditional or structured data has an established set of tools readily available, artificial intelligence and machine learning are emerging areas and are enabling brands to collect and operationalize unstructured data or previously hard to collect data. This could be via natural language processing, optical character recognition, text mining or similar approaches.”
As far back as 2018, a report from SmartInsights revealed that 55% of brands are either currently using or actively investigating some form of AI initiative within their marketing practices. The rise of AI-enhanced applications will enable brands to make better use of the data they acquire, as well as for marketers to show the value of their efforts. Once it has been properly “trained,” AI-based marketing will enable marketers to quickly and efficiently analyze large amounts of data from email, social media, websites, service inquiries, and other sources. “Remember that AI has to be taught and is only as useful as it is well-taught, which requires a data scientist. You have to get to a point where your AI can pass its own ‘final exam,’ so to say, before you can use it,” said McAdams.
Corugedo told CMSWire that even with the growing complexity of delivering personalized customer experiences, technology exists that can help fulfill the promised lift of segment-of-one marketing. “The right technology, combined with strategy, process and change management overhauls, makes it possible to embrace complexity and realize the benefits of becoming a customer-centric organization.” Corugedo said that a pro-active, “process everything” approach is required in order to keep pace with the overabundance of data, but it needs to be processed in milliseconds. “If it’s not processed and presented for use immediately, it will sit unused and clutter your data lake because overwhelmed data scientists cannot keep up with the onslaught,” he said.
Digital CX platforms take this approach and embed automated machine learning, a real-time decisioning engine, Corugedo said, and intelligent orchestration capabilities, all of which are used to tackle the inherent complexity while providing marketers with a single tool with which to meet customers’ incredibly high expectations for CX.
Use Data To Answer the Big Questions, Not Every Question
While there are many technologies that can be used throughout the data lifecycle, McAdams cautioned brands that they must answer the questions around what they want to do with the data: how they want to use it, analyze it, report on it, display it, etc., before examining those technologies.
When it comes down to it, data must deliver a positive business outcome such as revenue, profit, customer value, churn reduction or some such combination of these metrics. It’s not enough to simply use data to answer questions, but rather brands need to determine which questions need to be answered. Kharidhi reminds brands that each piece of data must add incremental value to the brand. “It’s important for brands to evaluate both from a quantitative and qualitative standpoint. Is the data adding more predictive power? Is the data adding more descriptive power? Is the source reliable or is the data of high quality?”
Some of the questions that brands should seek to answer are focused on how customers interact with the brand, how high the brand’s name recognition is, how much of the population are being reached, and how people are using the brand’s data. “You also want to know what tracks back to revenue, do people understand your value and do they choose you over other sources,” said McAdams. “You want your customers to hold every aspect of your brand in high esteem.”
DataOps to the Rescue
The DataOps approach uses Agile software development principles and automated testing, containerization, orchestration and monitoring to increase the production speed of data pipelines for Business Intelligence (BI) and data analytics. It also means breaking down silos across teams and departments, and encourages collaboration with data specialists in the organization. Kharidhi told CMSWire that as the volume of data continues to grow exponentially and the types of data available get increasingly diverse, brands require an Agile approach to streamline their data processes. “DataOps enables increased speeds of data ingestion, transformation, governance, lineage and quality while also creating much tighter alignment between data scientists and technology professionals.”
Joe DosSantos, chief data officer at Qlik, a data analytics and business intelligence platform provider, spoke with CMSWire about how brands can use a DataOps framework to tame the data beast. DosSantos recognized that not all data is created equal, nor does it have equal value. He said that only a small subset of the data that is collected will drive contextual insights that have an impact on decision making and that the key is identifying that data and turning it into actionable insights at the time when decisions are made.
“This is the process part of a DataOps framework, where alignment on people, process and technology helps an organization effectively access, transform and deliver data across the organization for decision making,” DosSantos said.
DosSantos said that fortunately for brands, the technology to support the equation is available now to combat data overload. Data catalogs can be used to serve as a single repository for the relevant data, and can be used to help generate a greater understanding of what assets are available, with a goal of focusing analytics efforts on streamlining the right data for the right outcome.
“Creating alignment is essential to ensuring the ongoing success of even the most robust data strategy,” he explained. “Where organizations run into trouble is in allowing data strategy to deviate from business goals. When they’re misaligned, even an organization that has mature data discovery processes can find that their insights are no longer accurate or even useful.”
With so much omnichannel data coming down a brand’s pipelines every second of every day, many brands are beginning to drown in data. By creating a data strategy for obtaining good data, putting technology to work, using data to answer the larger, more important questions, and using a DataOps approach, brands can learn to effectively swim in their data.