Every company is a data company, or so the saying goes. Businesses of varying sizes and across industries have differing needs and levels of data analysis proficiency. However, many understand the importance of data in their businesses — they just aren’t adept at using it.
An all-or-nothing mindset is the first challenge to overcome. Not every company is going to be, or needs to be, at Spotify’s level of data use(opens in new tab). Spotify and companies like it, with their millions of customers and many products, are natural candidates for big data and artificial intelligence use. There’s a misconception that only bigger companies can effectively use data, but don’t let an aspiration for perfect analysis be the enemy of good data use. In this post, we will outline three major stages in the data-use journey and identify the junctures between the stages.
The Case for Leveraging Data at Businesses of All Sizes
For smaller businesses, the how and why of data collection and usage is often not obvious. But the size and resources of a business should not impact its use of data, said Glenn Hopper, CFO of Sandline, a litigation support and eDiscovery provider.
“It’s very easy to see how the biggest businesses use data and how they're training these machine learning algorithms to make recommendations, keep customers on websites and tailor the customer experience based on all this data they have,” said Hopper. “While smaller firms may never have the resources of those mega companies, they can emulate their offerings and make data-based decisions through smart use of back-office technology.”
In his book “Deep Finance: Corporate Finance in the Information Age,” Hopper notes an unexpected personal example: his work as the CFO of a car wash. Prior to him joining, an outdated accounting system was the most advanced business technology that the mom-and-pop shop used. Each location ran like a separate business, and nothing was standardized.
Hopper brought in a developer to build a platform that consolidated data from all of the chain’s locations. The team could see how many cars were being washed and where, then staff appropriately. Eventually, the tracking became granular enough to detail how much chemical solution was used on each car, the tunnel speed, spend per car washed and more. With that data, Hopper and his team were able to identify best practices and standardize across locations.
Eventually, a private equity firm took the car wash chain public, thanks largely to the chain’s use of data making it more appealing to potential buyers.
The example demonstrates how data use can benefit companies in unexpected ways. Particularly for businesses looking to scale, it is crucial to both get a handle on your data and use it well.
Advantages of Tracking Data in Smaller Businesses:
- Deriving business insights from data analysis, leading to smart cost cuts
- Cutting costs without impacting product quality, which increases profitability
- Creating the operational efficiencies of a larger business through data insights
- Allowing for better, more accurate reporting
- Increasing business value in M&A situations
Stage 1: Defining KPIs and Collecting Relevant Data
For companies that have not yet started their data journey, the first stage calls for identifying your business’s KPIs.
“I have worked in the startup space, and in the very early days, you're still defining what your product is and what exactly you're doing,” said Hopper. “And so KPIs kind of get thrown out of the window. But that can be a problem because if you're just frantically pivoting and trying to meet each individual client need with a one-off of your product, it will be a disaster, and it's never going to be scalable.”
Choosing the right KPIs(opens in new tab) involves determining what is important to your business by asking questions like:
- What is our company’s vision? Mission? Strategy?
- What are our strengths, weaknesses, opportunities and threats?
- How do we want to engage with customers?
- Who will be our key partners, and what will they expect from us?
If your business, strategy or landscape changes, then your KPIs likely should change as well, making it critical to reassess them at least annually for relevance.
As an example of a business just starting its data journey, Hopper offers the food truck. Owners can start in the back office, tracking quantities of materials purchased by weight, date and spoilage. In doing so, they might discover that they bought 10 pounds of potatoes on a Tuesday and threw away half a pound on the following Sunday due to spoilage. But what if they bought five pounds on Friday ahead of their busiest time? Could that get them through the weekend without spoilage? On the front end, the same food truck could track where it parked at what time, the weather conditions and what purchases came from foot traffic versus drivers. Simply plugging this data into a spreadsheet or a cloud-based database tool and analyzing the findings can help a food truck cut costs and increase profitability significantly.
Small businesses often make decisions based on anecdotal information and experience. However, using data could give them a leg up in optimizing their operations, which is now critical as inflation squeezes margins. For instance, an electrician might start by looking at its inventory: all the wire, conduits, switches and circuit breakers. How much of a discount could it get by buying in higher quantities? Did each truck have the materials on hand to complete the job, or did the crew need to leave the jobsite to get items it didn’t have, and how did that affect the time to complete the job? From a geographic and scheduling standpoint, how does covering greater distance impact vehicle maintenance and gas costs? Is it worth it to take a job outside of the normal service radius?
Ultimately, it comes down to using the tools you have available to identify and track relevant data points.
“It’s getting so easy now to track data [given all the technologies available]. You just need to use the tools you have at hand,” said Hopper. “I think in any business situation, we could come up with a minimum of 10-15 data points that we could collect just from the phone in our pocket or, ideally, a business device like a point-of-sale system. For instance, a business using a payment processor could go to their site for reports.”
Next Steps: Determining where to log the data is likely the biggest hurdle businesses will face at this stage. While manual tracking is better than nothing, it can prove difficult and cumbersome, especially as the business grows. After companies identify their KPIs and where the relevant data is coming from, they need to start thinking about which system they will put in place to collect the needed information.
Stage 2: Consolidating and Finding Ways to Analyze Data
At this point in the data journey, companies have already decided on the data they want to track and how to log it. In the second major stage, businesses consolidate their data and find ways to analyze it – a step that Hopper is currently taking with his employer.
Sandline has an accounting system, project management system, a CRM and an industry-specific tool that its professionals use for eDiscovery and litigation support work. All of these tools have useful data – getting to it easily and getting it in one place is Hopper’s current challenge.
“Right now, we are extracting data from the project management system and the operational system, putting it into a comma-separated value (CSV) file and then uploading it into the accounting system,” said Hopper. “But we are actually in the middle of NetSuite migration right now to consolidate all this information. It is all going to be in the same system which makes things easier for me since we won’t be pulling this data from different areas.”
Syncing and consolidating allows you to start analyzing the data and eventually to start using more predictive analytics through making correlations. For instance, monthly recurring revenue metrics may tell you that you lost a number of customers recently. However, you won’t know why until you bring in other sources of data like exit survey results or full customer profiles. From there, you can start understanding why customers are leaving and mitigate churn factors to prevent future loss.
Top Tips for Data Syncing and Consolidation:
- Choose a system that can house various data sources and types in one place. This decreases manual work and increases real-time visibility.
- Map out current data sources and flow to understand your full inventory of data and usage.
- Delete outdated or erroneous data. Actively managing data during the consolidation process helps prevent an overload of irrelevant and/or incorrect information.
- Standardize your data. Go through all existing systems to ensure you use standard formatting and data types that can be combined into a single dataset.
- Ensure data completeness. Fill out all required fields prior to merging.
In stage 2, consider investing in tools like an ERP system(opens in new tab) to sync and consolidate data, as well as pull together dashboards with your selected KPIs.
“Through the implementation of a robust ERP system and commitment to gathering and using client data, even midsize firms can leverage data to provide robust business intelligence platforms that add greater value to their clients – just like the Fortune 1,000 companies,” said Hopper.
There are valuable insights to glean from an off-the-shelf system. Even the most basic set of canned reports allow you to run variance and exception reports, compare actuals to budget or actuals to prior period, and even analyze important factors like seasonality.
While you are getting comfortable with regular data analysis, this stage does not include hiring a data scientist. Instead, it is an analytical team member stepping up to the plate for data-focused activities.
Next Steps: The phase between implementing an off-the-shelf system with data analysis capabilities and hiring a full data science team – which we will get to shortly – is having someone in the company responsible for business intelligence using a BI tool to collect, organize and analyze data. The goal is to find correlations that are perhaps very specific to your business or weren’t completely visible in the canned reports.
Catalyzed by larger companies’ demands, the majority of data software, services and other repositories have made this process simpler by developing interfaces that allow access to data for analysis.
As you advance through the latter part of this stage, according to Hopper, you will ideally begin using application programming interfaces (APIs) or use open database connectivity (ODBC) connections to access data and analyze it. You will start writing SQL or JSON queries to get the analysis your company needs. While using these approaches requires programming skills, there are typically documentation and user-contributed examples on public forums like GitHub to get your data analysis expert started. At this advanced stage, you will start exploring tools like data lakes and data warehouses, as well as implementing systems for real-time replication of data.
| What Is a Data Lake?(opens in new tab) | What Is a Data Warehouse?(opens in new tab) |
| Stores a vast pool of raw, unstructured data for which a purpose hasn't yet been defined. Data is selected and organized when needed. | Stores structured, filtered data that has been processed for a specific purpose. Analysis is done on the cleansed data in the warehouse. Business intelligence systems often make use of a data warehouse. |
Stage 3: Refining Algorithms and Moving Toward Real-Time Analysis
Congratulations! If you hit stage 3, you’ve progressed further in the data journey than the majority of businesses. At this stage, your company is ready to become a data science-driven organization. This also marks the point where the company can level up and invest in an actual data science team that works across departments to find and reduce or eliminate inefficiencies. Tasks for the team can include:
- Helping marketing improve conversion from the website through tactics like A/B testing and heat mapping.
- Identifying how customers use a product or service and which services/features see the most engagement.
- Determining where and why sales might decline.
- Deciphering the most common types of support calls and building chatbots or FAQs to lower call volume.
There are multiple ways to set up a data science team, but Hopper likes the idea of embedded data scientists in each department, with a central employee or direct reporting line up to senior management. This data science strategy which Target famously used(opens in new tab) ensures that departments aren’t siloed and data scientists aren’t beholden to one particular group.
“I like the idea of an embedded team where you can have data scientists in each department that are still reporting to a central group,” said Hopper. “By embedding them, they really can understand the business needs. And then it's easier for the team that's going to be doing the development to prioritize workflow and provide value to each department and the business on the whole.”
Next Steps: While you may be nearing the end of the proverbial data journey, there is always room for improvement, particularly as new, more advanced technologies come out. Data science teams should focus on continuous improvement by adding new, more complex tactics like Hadoop clusters, refining algorithms and improving predictive analytics. We saw Netflix offer a prize(opens in new tab) to anyone who could beat its algorithm for predicting user ratings on films, creatively outsourcing part of this continuing task.
The Bottom Line
Building a data-driven business is a journey. Your data use doesn’t have to be the most advanced, but you shouldn’t dismiss data analysis as reserved for the biggest of companies. Regardless of size, resources or industry, businesses have an opportunity to benefit from data-driven insights at a scale that works for them.