Data: Fuel for Growth
Best practices for companies looking to integrate big data into their operations
As the value of big data and the power of sophisticated data analytics capabilities become more evident, these tools are moving from the category of a luxury to a strategic imperative for successful businesses. A rapidly expanding set of data sources, consistently improving tools, and advanced data science methods have made these tools and techniques more accessible. Today, it is not just cutting edge startups and tech companies that are successfully harnessing the science behind digital business. Some of the best practitioners are in businesses that have been around for decades, and are integrating sound data science into their business models in ways that confer a significant competitive advantage.
While a big data approach is traditionally viewed as the province of business-to-consumer companies, innovative business-to-business companies are uncovering key insights about their customers and business processes by marshalling the data at their disposal. From retail to the automotive industry, data analytics is influencing customer engagement and shaping the way companies engage with their professional partners. It is also leading to new insights and contributing to more informed decision-making across a range of industries and a host of new applications.
But facilitating data collection, enhancement and analysis, and practicing decision science in a way that drives meaningful improvements in performance is not always easy. It requires an understanding of the best practices associated with implementing an organized and proactive data management strategy, and designing and deploying the technical architecture, professional talent and systems and software required to make that happen.
Successful adopters move quickly from data gathering and analysis to engage in rapid learning, early experimentation and iterative product development Assets and assessments
The progressive use of data science begins with a mindset shift—recognizing that all of the data your core business generates is an asset, and treating it accordingly.
The data available to you might be transactional data or customer data, or a whole host of different metrics and measurables surrounding performance and operations. But once decision-makers recognize the hidden value of that data, they can move to capture it. And they can develop the technical architecture and operational parameters and processes to collect, store and manage that data.
While the scale of investment can differ significantly from one company to another, the combination of a small team of in-house data scientists and a diverse set of external analytical and data science resources can be a potent one. The most aggressive and successful adopters move quickly from data gathering and analysis to engage in rapid learning, early experimentation and iterative product development. Open source data science tools, virtual team connectivity, cloud computing, and communities of young independent collaborators can all help accelerate and enhance that process.
Step by step
Adopting agile data science techniques at an enterprise level can serve as a vital part of wholesale digital transformation initiatives. Companies that recognize the value of their data may take proactive and even aggressive steps to assemble a world-class data science team and invest in new technology and capabilities.
Deciding to take the data analytics plunge is only half the battle, however. To be effective, you will need to proceed in a deliberate and strategic fashion, building a data management and analytics infrastructure using the following basic steps:
To be effective, your data analytics efforts need to be closely tied to a strategic pillar. You need a reason to do this beyond simply embracing something new and different. Start by identifying and articulating your strategy. Once that strategy is clearly defined, you can then ask yourself where you can leverage data in service of that strategy.
With your strategic premise established, move to conduct a comprehensive data audit/review, the goal of which is to get a clear sense of what your data assets look like. Remember that it’s not just what you have, but what you can use. Maybe you have detailed and extensive customer data, but are restricted from using it for various regulatory/legal reasons.
Once you have identified your usable data assets, focus on the key drivers for your business and your customers. The next step is to determine what (if any) external data sources you may need to augment your internal data resources.
With the data landscape and sourcing established, you can then move forward with developing the specifics of the overall strategy to put those new assets in motion. Be sure to rigorously test each step in the process, along with the ultimate outputs. Consider what kinds of business intelligence tools you may want to use to access and review your data and your conclusions.
Tools and technologies
My own experience in the automotive remarketing space provides a ready example of what this process looks like in action, and what kinds of tools can be developed as a result. The goal of our data science team is to use big data and analytic capabilities to streamline the remarketing experience and deliver clear, actionable information to sellers and buyers. We aggregate transactional, customer and market data to build predictive models and product algorithms. When paired with external economic and market trend data, and integrate it into the auction processes, these solutions help sellers and buyers make effective pricing and venue decisions and better predict sales, revenue and inventory.
This approach to data science also affords opportunities to develop new tools that further refine your models and enhance your data analytics capabilities. Our tools include:
• Computer vision. tools that enable machines to identify damage features from digital images and evaluate the condition of a vehicle. These tools have the potential for producing significant financial and operational efficiencies throughout the inspection process.
• Pricing models. that combine information about condition, make, model and trim with deep analytics to assess a vehicle’s true value in the current market and in forecasted market conditions. This helps avoid aspirational pricing and hone in on realistic market pricing now and in the future.
• Dealer-vehicle affinity. tools that assess which buyers tend to demand certain types of vehicles at optimum prices. If a dealer frequently pays more for a specific model because that model is in high demand in their market, this picks up on that.
• Venue selection. Another layer in the modeling evaluates which market venue, platform, and/or region should be used to optimize value and reach receptive dealers.
While the big data revolution is only just beginning, it is clear that pioneers in this space have an advantage. Companies that can successfully integrate big data into their operations will be able to streamline processes, generate new insights and efficiencies, enhance customer service, and ultimately make a meaningful difference to the bottom line.