In Data We Trust: But Do We?

In the last few months, we have seen data used, misused and abused to report on the COVID-19 pandemic. The only consistent opinion of the data that has been shared with the general public is that we’re not sure what to believe. Is the virus really that rampant? Do masks really make a difference? How many deaths were caused specifically by the virus?

Data has attempted to answer all those questions, but the answers are still not trusted. So, what difference does it make whether we trust data or not? Why is it important and how do we ensure that the data we use to run our businesses falls into that TRUSTED category.

In Stephen Covey’s book, The Speed of Trust, he makes the point that “trust impacts us 24/7, 365 days a year. It changes the quality of every present moment and alters the trajectory and outcome of every future moment of our lives.” But, Covey argues, the real reason that we need trust in business is because it contributes significantly to the speed at which we respond to our customers, our employees, and our boards. “When trust goes down, speed also goes down and costs go up,” he writes.

Here are two examples of how lack of trust in data effects the speed of your business.

1. The Revenue Conundrum: Your business manager has just handed you sales figures for the past quarter. You look at the bottom line and it doesn’t seem right. Last month, the revenue figures were off by $500K. So, immediately you eye this month’s figures with skepticism. You don’t believe – or trust – the data. Without hesitation you deploy additional resources from your team to “look into those figures.” A few days later the team has pulled together a new report that you think looks closer to what you think it should be. But your team is now days late on another project and the decision you would have made to help generate more revenue has yet to be put into action. And, finally, you have expended resources to do work that had already been done.
2. The AI Algorithm: You just finished reviewing the latest results of the Data Scientists’ team and the direction that was recommended is exciting. However, when the team explained what they had to go through to get the right data set, you wondered how the team has analyzed the last 5 years of resumes and hiring data from HR to recommend a model for hiring a more diverse employee population. When the project was tested, however, the model recommended to hire individuals that fit the same “thirty-something-white-male-Ivy-League-school profile.” The data that the team used was accurate and plentiful but biased toward continuing the same hiring practices.

Data integrity in today’s world is so important because AI algorithms learn from the datasets they initially train on. That learning becomes the basis for additional learning over time. If the initial dataset is not representative of the problem at hand, then the AI algorithm will produce spurious and ineffective information. In fact, this AI example comes from a high-tech company in Silicon Valley who spent 4 years and millions of dollars on a hiring model that they scrapped as a result of using the wrong data.

These are not isolated examples; these scenarios happen in businesses across the world – daily.

A recent NTT DATA study found that, while 78% of companies in North America understand the value of their data, only 43% of those surveyed agreed that data plays a pivotal role in the decision-making processes in the company. And both data leaders and laggards from the study conveyed that “senior leaders not seeing the value of using data” remains a challenge when attempting to derive meaningful value from their data.

So, how do you put trust back into your data? How do you get senior leaders to see the value in data and trust it?

Using Covey’s approach and examples in conjunction with insight from our study, we can look at our data in four ways that helps create credibility. What are your answers to these questions?

1. INTEGRITY: Does your data have integrity? Does it truly reflect what we know is reality? Is it current, complete, consistent and accurate to the best of our knowledge? How do we know the data is high quality? What good decisions have we made with the data presented to us? About a third (34.2%) of organizations identify improving data quality as biggest challenge in managing their data.
2. INTENT: What is the Intent of the data? Is the data collected and used in a way consistent with the intention of its use? Unfortunately, our study found that about a sixth (15%) of companies don’t understand what data they have. The promise of privacy is a good way to evaluate intent. Is the data you are using from reliable sources? Do you have the owner’s permission to use it? Are you using data collected for sales activities that are now being used for HR activities or vice versa? Does the data fit the intent?
3. CAPABILITY: Do you have capable data people? Do you have people who understand the meaning of the data that you use and depend on? Do they understand its lineage, its flow through the organizational applications and its maintenance? Can these people establish and/or restore trust with the users of the information the data provides? About a fifth or 21% of our study respondents don't have the necessary skills and talent to analyze the data.
4. RESULTS: When you use data, do you get the results you want? Does the data deliver quality information consistently? Can you quickly assimilate the information derived from the data and apply it to the business decision you need to make? Data gives you information that can be applied to the business. Are you getting value from your information? That’s why this step – RESULTS – is so important. Our study revealed that only 10% of the organizations use data very effectively for transformational processes: and 17% of respondents say they’re not using data effectively at all.

If you lacked answers or answered negatively to any of the questions above or were hesitant with your answers, you might have an issue of trust in your data. That lack of trust will slow you down and create additional expense in your organization. Here are some simple actions to take toward establishing trust in your data.

1. Assign a data owner. This owner should be given responsibility for the data, but also the power to act on its behalf. Although you can hire this individual from outside your company, it’s often best to select someone internally who has the trust of your employees and leaders. Trusting the data owner is a first major step to securing trust in the data. And don’t make the mistake of thinking that an executive is necessary for this task. We’ve seen many governance programs led by individuals pulled from throughout the organization. It all depends on the culture of your company. Who do people trust the most? A little over one-third or 37% of our survey respondent say that they have appointed owners for all their data sets.
2. Identify a key project. Identify a key project that links to or contributes to the overall goals of the company. For example, if your goal is to generate 10X the revenue from last year, then use the data to identify what products and channels you will need to expand to reach the goal. Use the data also to tell you where your weaknesses are in the supply chain, the go-to-market and financial operations. Make sure that you always start with the data associated with the process or task. Give it a spotlight so that, when you reach your goal, data can be a key contributor.
3. Develop systems, processes and people. Only 31% of respondents from our survey say that data is shared seamlessly across the enterprise. This is the big step. You may have systems, processes and people already, but are they tuned to the requirements that you have for a data-driven environment? Do you have applications whose data does not link to or talk to other applications? Do you have processes that are broken or incomplete? Do you have people who understand the data from their area of responsibility, but haven’t a clue as to how their data effects the entire organizational process?

In summary, trust in data begins with identifying what is leading to the lack of trust, putting a plan in place to correct any issues, and then using data to make business decisions, accelerate growth, save costs and drive employee and customer satisfaction.

Get your copy of NTT DATA’s latest research, or contact us today for more insight on how can your organization can turn data into an asset or strengthen data governance to improve the trust in your data.