From cloud-based storage for personal photos to the expansive sensors and cameras used in self-driving cars, the proliferation of data applications means data scientists and consumers alike now benefit from more freedom and opportunities to use data every day. Storage and computational power are major inputs into how that data can be used, of course, but perhaps the most important piece of the data puzzle is the interface software that helps employees, customers, and businesses make the most of all that highly accessible data.
Given the unprecedented accessibility of useful data, companies should take note of the potential benefits of finding and using the right software. Doing so can help data science teams manage the creation of new data and facilitate the use of future data.
The Pros and Cons of Build vs. Buy
Deciding between paying licensing fees to use pre-built software platforms or investing in a team that will create custom software to solve issues specific to your company is difficult. The variables change between situations and organizations. Licensing software can shrink your long-term profit margins, as you’ll always be paying another company to use their software. Alternatively, the time required to build your own software and costs of staffing can front-load the expenses of getting your solution up and running.
Whether you’re considering a large-scale investment or an individual tool, compare the ROI for each scenario and keep your organization’s long-term goals in mind. A key question to ask yourself is whether the technology will still be useful in five or ten years. If not, perhaps leveraging another company’s software might allow you to shift to other software more easily, providing you a nimble and flexible environment wherein you can move to better software as it becomes available. If you build your own software and have heavy sunk costs into your own application development, this might reduce your technological flexibility in the future.
Alternatively, perhaps your core applications demand a high level of customization so you can cultivate a competitive advantage in your industry. If you’re using software that even your most aggressive competitors have full access and visibility to, competitive advantage might be more difficult to realize. Regardless of which option you choose for a given situation, finding the right software to support data science offers measurable benefits to any organization, so the decision to build or buy is a significant one.
A Forward-Thinking Strategy
The most important part of deciding to create your own software is challenging your team to think ahead. Contrary to what most people think, data scientists don’t just use existing data. An important part of their role is also managing the creation of new data. The future efficiency of our business at Clearlink and our opportunities for success depend on the type and amount of data we can access, generate, and store.
In other words, being able to create future opportunity depends on having a continuous supply of data. We can’t just pull data, organize it, analyze it, make a few changes, and then sit back. If we want to have real and lasting impact on our customer experiences, we always need to think about the data we will need to solve the next problem.
That being said, a business needs the right data to make real change and improvement. A set of generic data may illuminate a few areas for improvement, but it also may just reinforce what a department already knows are its pain points. Thinking ahead to what kind of data will be most useful to the problem at hand and building software that will be able to capture that data in the future is key.
Based on our experience, there are a few things your data science team should focus on:
- Continuously gather useful data.
- Effectively store and organize that data.
- Understand what additional data would be useful.
- Think about how data creation could help the organization accomplish more of its overall goals.
Of course, any data science team will have other unique goals to address at any given time, but these basic points are important to keep in mind. Building data-forward software works best with a proactive mindset, and these four actions ensure your team’s focus stays on the future and larger goals, rather than just day-to-day tasks.
Having Data That Works for You
We don’t have that data, so we can’t make that improvement.
This is a common refrain—but with the amount of tools available now, it’s no longer a suitable excuse. If your data is leaving you with gaps in understanding, don’t let it limit your team’s progress. Figure out if there’s an existing tool you can buy to get that information, or if you have the resources to build your own to get the information that solves your problem.
At Clearlink, we’ve been gathering data to help sales agents personalize conversations and recommendations for customers. To build a large enough supply of data to be effective, our data science team is building software that gathers and stores data from conversations in real time. When a customer tells a sales agent they have a dog or mentions that they’re interested in photography, for example, our data software will record those bits of information and then make that information available to the sales agent the next time that customer calls.
Suddenly, it’s possible to make personal recommendations for a cable package that includes a dog channel or an add-on camera sale based on information from past conversations. The more information we can capture about a customer, the better equipped we’ll be to create an improved experience for them the next time they call in.
Software That Creates Solutions
Clearlink has invested in building our own data processing software because it lets us capture more data on our own terms and increase our opportunities for efficiency. That being said, we want to make sure we’re creating useful data that encourages real efficiency. The software we’ve built to track customer calls and store data has enabled us to launch a brand new customer care content department as a solution to a variety of common customer needs.
By looking at service calls and tracking the issues customers are calling in about, we can create content that caters to more specific, but still common, concerns. This goes beyond providing basic information, and it isn’t meant to address customers who prefer to use a phone over a website—we want to help customers who are looking for solutions on our website but can’t find them. If many customers call with questions about specific areas of their bill, we tweak our website to make sure it’s answering those common questions. We’re sorting through each customer call we get to do this for every resolvable issue.
In a large call center, margins are often thin. Deflecting even 20% of calls can reduce operating expenses by 20%, which translates directly to increased profit. That not only serves the business well, but it serves the customer better too. Nobody wants to be on the phone for twenty minutes if they don’t need to be. Getting customers the answers they need faster benefits everyone.
We’ve also written software to centralize our data and make it available in real time to all of our applications. This enables us to pass along information about where a customer has been on our website before they reach the call center. Several years ago, a lack of technology and prohibitive costs would have made that impossible. Without those limitations, we can capture, store, and deploy important information to our agents and solve customers’ needs faster—even on calls we’re not able to divert through our customer care content and other deflection efforts.
When we considered buying software versus building our own to improve the customer experience, we ultimately chose to create our own because we wanted a platform that served our business’s unique needs. We didn’t want to pay for features we weren’t going to use, and we didn’t want to spend time customizing a generic platform when we knew exactly what we needed from the start. In this case, the investment in personnel and time that was required to get a project of this scale off the ground still made sense because this software is answering long-term business needs and directly affecting profit.
As the cost of data storage drops and the technology evolves, investing in data-forward software makes sense for most companies. Once an internal data science team understands pain points and opportunities for improvement, integrating data needs into software development sets the department up to make more positive changes in the future. Looking ahead is paramount to data science, and writing your own software can be another important part of staying proactive and being in control of your company’s future.