My last article discussed three key decisions you should consider when building an analytics unit in your organization. In this two-part post, I want to dive deeper into that topic. I’ll unpack the keys to building an analytics team for your organization in this article.
Grab a notebook and a cup of coffee because there are a ton of golden nuggets for you inside. Part 1 of this article breaks down into the following sections:
- Use cases
- Infrastructure
- Staffing
- Budget
Part II will cover the following additional sections:
- Organizational Structure
- Governance
- Analytic Processes
- Cultural Change Management
Use cases
You’ll need to start by thinking about the use cases for your analytics team. Use case refers to an individual problem or question you want analytics to address.
If you aren’t familiar with identifying use cases, consider each of your business’s functional units first. For each unit, think about your specific questions or problems where analytics could shed light or provide an answer.
You can brainstorm this list at the beginning. Don’t worry about the feasibility of each use case right now, but make sure you are being specific. You’ll shorten the list in the next step. This article teaches you more about asking great questions.
Next, you’ll review the list of use cases to see if you have the necessary data. For each use case, identify the two or three most essential pieces of information to describe the issue. As with the use cases themselves, you want to be specific about these two or three pieces of data.
For example, do you need to know:
- Revenue by customer?
- Did a customer purchase a specific product or service?
- Specific demographic characteristics of the customer?
If you identify use cases for which you don’t have any data, move those onto a list of nice-to-haves. You can tackle these use cases down the road. When building your analytics team, focus on their work at the outset.
Staffing
With your use cases identified, you will need to identify the number of staff and skills required for your analytics team. The number of staff will be determined by the volume of work you want to complete, and the skills needed by your analytics staff will depend on the complexity of the work.
Entry-level analysts should suffice if your use cases consist of basic descriptive analysis of your production and sales databases. In contrast, a higher-level data scientist is warranted if your use cases consist of complex modeling and segmentation analyses.
Your staffing decisions will also depend on who analytics reports to and the organizational structure chosen. If your new team consists of entry-level analysts, they will need a manager familiar enough with the use cases to guide their development. In contrast, if the team includes more senior data scientists, they may be allowed more autonomy from supervision.
Depending on organizational size and use cases, a startup analytics team will likely have between one and five staff members. Additionally, unless you have leadership with analytical skill sets, you probably won’t want to hire entry-level staff initially. If you start with senior staff first, you can backfill a team around them in the future.
I’ll discuss your choice of organizational structure in detail in Part II of this article.
Infrastructure
You’ll need to identify the critical pieces of infrastructure your analytics team will need to achieve success. In this case, the infrastructure you should pay attention to consists of:
- Computer hardware
- Software licenses
- Online services
- Third-party data purchases
Computer Hardware
One obvious infrastructure requirement is computer equipment. You won’t get very far in analytics without it. At the minimum, each team member will need a laptop.
Data security requirements vary from one industry to another and across geographical regions. Depending on your business and location, you may need to invest in infrastructure to increase data security. Such infrastructure could include physical server space to house user data and software. However, cloud server solutions are increasingly meeting increased security demands.
Software Licenses
Depending on your analytics team’s work, they may need specialized software. For basic productivity, you’ll likely want at least Microsoft Office or a Google Workspace license.
For analytics, you must decide whether to use open-source software like Python or R or a paid platform like SAS, Stata, or SPSS. Additionally, consider options such as Microsoft’s Power BI, Tableau, or other platforms if you want data visualization.
I recommend seeking outside counsel if you are uncomfortable making hardware and software decisions. Find someone in your network with this experience and tell them about your business. Ask them what configurations they would use if it were their business. Personal preferences sometimes drive these decisions, so you might get different answers if you ask multiple sources.
Online Services
The subject of online services overlaps with software if you’re using Google Workspace for productivity or other Software-as-a-Service (SaaS) tools. However, I am referring to online services regarding data storage, processing, and integration.
Today, you can set up an entire business IT infrastructure in the cloud if necessary. Tools such as Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP) offer a wide array of data storage, processing, analytics, integration, and security capabilities.
In addition to these full-service options, numerous online services provide a piecemeal approach. Tools such as Google Drive and Dropbox offer easy document storage, while platforms such as Snowflake, Databricks, Mozart, and Panoply offer cloud data warehousing.
Services such as Saturn Cloud, Deepnote, and Altair Rapidminer provide cloud-based data science services with robust built-in workflow, analysis, and visualization capabilities. Many analytics platforms offer Python-based analyses and AI assistance.
Services like Make and Zapier provide drag-and-drop workflow development to help integrate platform processes. These platforms are commonly used for syncing and merging data captured across multiple services using real-time triggers and automation.
Other online services may be helpful depending on your business. With the sheer number of services coming out daily, many options exist to work smarter instead of harder.
Third-Party Data Purchases
In the world of analytics, data is king. An entire industry focuses on making captured data available to commercial clients. Some of the best-known third-party data providers are Nielsen (television audiences), CoreLogic (property), and Experian (consumer credit).
You might consider third-party sources if you don’t have proprietary data for all the factors in your analysis. However, be aware that third-party data tends to be expensive. After all, a company collects, cleans, and compiles that data on a large scale for you.
Additionally, third-party data often has limitations you’ll want to consider. For example, you may not be able to link data at the individual level. Or, based on the process used to collect it, the data may only cover part of the population.
Remember that third-party data is often available from your city, county, or state government. For example, your local Department of Revenue might have data on newly formed businesses. Similarly, the county assessor’s office might be able to provide data on local property transactions.
Importantly, please keep in mind that even government data sources will tend to have a processing cost. Additionally, while these third-party sources are often cheaper than commercial data, those savings come at a price. You will probably need far more work to clean and prepare the data for analysis than a commercial third-party source.
Budget
I’ll start this section with a divisive idea. Many business owners begin by asking whether they have the budget for hiring first. I’ll encourage you to make your budget review and considerations last in this process. If you scope out the position description and staffing requirements first, you can make a more informed budget decision. You can also adjust the hiring scenario if you make the budget considerations last.
If you google the cost of hiring employees, you’ll find estimates from $4,000 to $30,000, excluding salary and benefits. If you include compensation and benefits, the cost estimates are around 1.2x to 1.5x annual salary.
The cost of hiring an employee will vary based on the tools used, whether you use a staffing agency, the industry, geographic location, and even the type of position. That being said, your budget considerations for hiring an employee should include factors beyond the cost of hiring.
Do you have a sufficient overhead budget and a financial buffer? It would be unfortunate to hire your first analyst right before a revenue downturn, but it happens. Be sure to include a review of your revenue projections and capital reserves as part of the budgeting process.
There are alternatives, even if your budget doesn’t allow a full-time hire. You might be able to contract analytics to an outside vendor or find a part-time analyst to complete specific projects.
End of Part I
That’s plenty to chew on in one sitting, but we’re only halfway through! I’ll pick up again next week with additional considerations for organizational structures, governance, analytic processes, and cultural change management.
Until then, I wish you all a great day and a successful week!
