For many organizations looking to transform into data-driven firms, the largest roadblocks often aren’t technical but cultural. While the process of analyzing and deriving insights from data to inform decision-making has become mainstream, the difference between data-driven firms and those struggling with data often boils down to culture, as per a McKinsey report.
Companies need to inculcate a thriving data culture across functions — including business operations, product, marketing and human resources — so they are equipped to make informed decisions. This means embracing a fundamental shift in mentality.
As a company’s multiple departments become stronger at addressing and extracting at least basic data queries and visualizations, its data analysts and scientists will get the space they need to focus on their core responsibilities—improving data models to provide meaningful analyzes beyond simple data pulls. In other words, developing data intelligence support systems to ensure timely availability of data services.
These data services come in different forms. The high-level metadata categories include:
- Behavioral: Keeps track of who uses the data and how
- Technical: Displays the definitions of a schema or table
- Business: Policies on how to properly manage the various types of data
When a new version of a dataset is developed, the provenance shows the relationship between two versions of data items (also known as lineage.) Behavioral metadata is extremely significant since it represents an organization’s human expertise around data. It demonstrates how individuals can use data to derive insights and learn. The animating spirit of an organization’s unique data intelligence is formed based on how employees use data.
Organizations would do well to adopt the following approaches to building a strong data culture:
A committed top management
Data-driven culture in any organization should begin at the top. The senior-most executives and managers need to establish that decision-making, for business-critical moves at least, need to be based on data, and that this approach should be the standard, not the exception.
Choose metrics with caution and ingenuity
Leaders can significantly impact behavior by selecting what metrics to monitor and what they expect their staff to use. If a company can benefit from predicting pricing changes by competitors, it should have a team focusing on regularly drawing up specific forecasts regarding the amount and direction of such shifts. It should also keep track of the accuracy of the forecasts, which will improve with time.
Data scientists in management
Data scientists should be hired from within a corporation to leverage their knowledge gap understanding. If the analytics team operates separately from the rest of the company, it will find it difficult to create value. Companies can address this challenge by ensuring that its data scientists either:
- Have line management experience, or
- Are given line management training and experience.
Accuracy of data
Decision-makers need to confront possible sources of uncertainty head-on: Is the information accurate? Is it possible that there aren’t enough instances for a valid model? How can elements like growing competitive dynamics be included when there are no relevant data? To understand the importance of this, consider a supermarket chain. If the inventory data is inaccurate, the procurement department will be misinformed, leading to supply problems and a fall in revenue.
To prevent this chain, data scientists should be able to get accurate inventory data in time so the procurement team can take appropriate actions based on that data. This requires data scientists to check for accuracy on a recurring basis to ensure that the data generated is reliable and actionable.
Specialized and focused training
Specialized training should be provided only when needed. Many businesses spend on bang-for-buck training, only to have their employees forget what they’ve learned if they don’t put that to use straight away. While skills such as coding should be included in basic training, it is more beneficial to teach employees specific analytical tools and ideas such as proof of concept only when required.
Consider a shelf stocker in a retail store who now needs to fill in as a cashier. The training for the cashier position should be given just in time and should be based on data. This way, the training will be fresh, and the job performance efficient.
Quickly resolve data-access concerns
Data access should make data easy to use. Only data that’s useful or necessary to a particular employee should be made available to them. If universal data is shared with all employees, the information might prove complicated for many, and they might not use it at all. On the other hand, if relevant data is shared with employees in an understandable format, it is likely to be used more frequently and meaningfully.
A task becomes a choice if it benefits an employee directly—such as saving time, avoiding rework, or retrieving often required information. For example, sales personnel will not have use for accounts payable data. But give them customer data in an easy-to-understand format, and they will likely put the data to great use, making for a data-oriented culture.
Trade flexibility for stability
Be prepared to give up flexibility in exchange for stability. Many companies that rely on data for their decision-making have a variety of data categories. But if each employee has their preferred information sources, metrics and programming languages, that would be disastrous for the company. Trying to harmonize somewhat different versions of a measure that should be universal can consume much valuable time.
Also, inconsistencies in how different modelers work have several implications. If a company’s coding standards and languages differ, for example, every move by the analytics personnel would necessitate retraining. Internally sharing ideas might also be excessively time-consuming if that continually requires translation. Instead, businesses should use canonical measurements and computer languages.
Alternatives appear too dangerous. Companies—and the divisions and individuals who make them up—frequently fall back on habit. Data may be used as evidence to support assumptions, providing managers with the confidence to venture into new areas and procedures without incurring risk. Merely wishing to be data-driven is insufficient. Companies must create cultures that allow this attitude to flourish. Leaders can foster this transition by setting an example, modeling new behaviors, and setting expectations for what it means to make data-driven choices.
Get employees excited about data
If employees are given general training on data and data use, they could get bored and leave the task altogether. But if companies can link the training to the immediate goals that each department needs to achieve by using data-driven techniques, they could get their employees excited about achieving those targets.
For example, if the sales department is taught to use data to shortlist clients, they can use the technique to achieve short-term goals, and will be excited to learn in the future as well.
Because each organization’s culture is different, you’ll need to create a custom stack of solutions for possible inconsistencies due to the requirement of the finance department’s buy-in on new investments. That is why it is critical to tackle it correctly.
As a start, make a list of the solutions you’re utilizing and see whether what’s working for one business unit can be extended to another. This entails adapting solutions to new use cases within the company for many software firms. That won’t always work, and you’ll probably need to invest in new tools as well, and so you should audit your organization’s goals and create a framework based on what you learn.
While this is a lot of additional responsibility, it allows departments leads and teams to work closely with one another, encouraging innovation and deeper business insights. After squeezing years of digital change into mere months because of the pandemic and subsequent lockdowns, IT executives have a chance to emerge as equal partners in creating cross-departmental vibrancy and development.
Conclusion
While corporations have always been interested in their numbers, in a data-driven culture, the degree of data utilization is exercised at a higher level. The major goal is to enable all employees to actively use data to improve their everyday work and to fully maximize a company’s potential by creating decisions that are more successful, projects that are more effective, and competitive advantages that are more obvious. One way to do this is by enlisting the services of an experienced partner like SAP. With the help of SAP data services or by using SAP data intelligence, support your organization in its journey to mine wealth from numbers.