Analytics are at the core of many modern business decisions—from large-scale mergers and acquisitions, down to customizing a marketing email to an individual customer, and everything in between. Unfortunately, it is far too common for companies to get caught up in finding a quick solution to one small problem or initiative, or in chasing the latest technology without considering how a new tool or process fits into their business goals and vision. This can lead to a “Frankenstein” data and analytics infrastructure made up of many disjointed parts that do not work well together.  

Each part may serve its own small purpose well, but when a business needs to consolidate its resources or when an opportunity arises to merge technologies for a new strategic initiative, the costs and challenges of doing so can be very high. These costs and challenges lead to unnecessary work and missed opportunities in the short term, and diminished business success later. 

When considering the implementation of a new technology or a new analytics project into an existing data and analytics infrastructure, it is essential to “keep the big picture in mind,” ensuring that the new technology or project works seamlessly with existing processes to advance the business’s goals.  

Don’t Fall for “Shiny Object Syndrome” 

In a rapidly evolving data and analytics landscape, it is easy to feel like your company is being left behind if you are not implementing the latest disruptive technology. There is perhaps no better example of “shiny object syndrome” than with the mad rush to implement Generative Artificial Intelligence (Gen AI) technologies.  

There is certainly merit in staying on top of the latest technologies to avoid being at a disadvantage in the marketplace. However, investing vast time and resources into new products and tools before they reach a critical state of maturity can often lead to unsuccessful implementations and high costs associated with adapting the project, as the technology inevitably evolves rapidly after its initial release.   

Businesses are recognizing this as well. A recent survey has shown that a majority of pharma companies are taking a very cautious approach to implementing AI solutions, citing security concerns as the technology matures, while others have called the technology overrated.  While Clarkston has had some great successes implementing AI for some of our clients, we also take a cautious approach, making sure that the technology is the right fit for the business’s goals. 

Businesses and project managers should exercise caution when pursuing an exciting new technology, ensuring that it will fit well within their existing infrastructure and that it aligns with their overall business strategy. 

Foster Technological Alignment 

Many companies centralize all their digital, data, and analytics operations through platforms such as Microsoft Azure or Amazon Web Services. These platforms are extremely flexible and can accommodate other tools like Marketing Technology (MarTech), coding environments, data warehousing, and many others. 

For example, in recent years, the cosmetics retailer Sephora has launched a number of new initiatives to enhance customer experience, implement Augmented Reality using AI, and make quality product recommendations based on carefully tracked customer data.  Critically, “Sephora doesn’t treat its website, app, and stores as separate entities. Instead, it has created an integrated ecosystem where each channel complements the others.” 

When company decision-makers, strategists, and project managers push a new initiative—say, implementing a new email marketing system—it is critical to choose a product that can be seamlessly integrated into the company’s existing data and analytics ecosystem.  Failure to foster technological alignment may lead to the use of numerous products and services, each with their own subscription costs and maintenance requirements that require separate processes to bring data into a usable format and location for other data and analytics tasks. 

Write Flexible Code 

When writing code for software development, data engineering, or data science, it is not only best practice to modularize the code, but it can also save time and resources as project requirements evolve. With modularized code, some components such as user-defined functions or model architectures may be used in multiple applications where appropriate, eliminating the need to rewrite basic components.   

Modularized code also simplifies script maintenance. As input data sets change or as models need to be re-trained and tested, it is far more efficient to edit code and run tests in shorter scripts than to run scripts in their entirety to test each small change. It is also ideal—though sometimes difficult—to write code for machine learning and artificial intelligence models that is flexible enough to allow for new features to be incorporated as new data sources are acquired or old ones are deprecated. 

In his book, Data and Analytics Strategy for Business, Simon Asplen-Taylor emphasizes the importance of creating processes that can be repeated to make the best use of resources.  Especially early in a data and analytics journey or early in a project lifecycle, he encourages thinking about scalability and progressing through a project in a way “that allows you to repeat your win in other business units, or more locations,” or even in other places within the same project. 

In a perfect world, a comprehensive end goal or vision for a project and the steps to reach that goal would be clear when the project is initiated, but that is often not the case. For that reason, flexible code allows for adaptability as projects hit roadblocks or as new ideas are presented. 

Most importantly, always ask yourself, “Does this simply solve a short-term need, or does it cohesively move the company toward a long-term vision?” 

Humans tend to seek instant gratification at the expense of long-term best interests, and that is certainly true when it comes to initiating analytics projects or implementing new technologies.  It is often tempting to launch a project that provides a quick fix to an immediate need without considering how the project may factor into the bigger vision. 

Anyone managing multiple data and analytics projects should ask themselves different versions of the question, “Does this advance our company toward our long-term goals?  Does this new technology align with our current data ecosystem? Will this predictive model still be relevant to our strategy in one, two, or more years? Can the insights we gain from this project be integrated with existing projects to gain even greater insights, and if so, how should we build it to align all projects as easily as possible?” 

Clarkston Consulting’s Data and Analytics team excels at prioritizing urgent business needs while helping our client partners stay on track to meet their ultimate objectives. If you’re looking for guidance on how to keep the big picture in mind when it comes to your data analytics needs, we can help.