If you landed on this page, you’re at least thinking about what you need to build analytic capacity. Congratulations! That means you are not just ready to succeed but to take your team to next-level performance.

To build analytic capacity, you’re going to need to nail down 14 key elements. Each of these elements will contribute to your success. However, don’t worry that you need to spend tons of resources and time in the process. Many of these items can be planned and executed faster than you think.

Each of the 14 elements is divided into five domains to organize your development. The order of domains is intended to be a logical sequence for developing your capacity. However, you can develop the five domains in any order you choose and still reach the same capabilities.

Use Cases

At the outset, you need to consider the specific reasons you want to build analytic capacity. Start by writing down your specific goals for building out your team to set goal posts for your effort.

Goals

When setting your goals, focus on developing the information you need to move the needle on building your business. You might want to improve your sales funnels with better personalization. Or you might want to learn more about the factors driving sales conversions. In these examples, achieving these goals would give you the information needed to improve your operations.

Specific Questions

In addition to defining your goals, you’ll want to list a few key questions under each goal. These questions should be more specific than the goals they relate to. Focus your questions on the most important aspects of the goal they sit under.

For example, to improve your sales funnel, you might test multiple messages to find out which one works best. To identify the factors driving conversions, you could compare customer characteristics across those who respond to an offer and those who do not.

Assumptions

Finally, when developing your use cases and the questions you want answered, consider any assumptions you’ve made along the way. For example, you may want to test your sales messaging. However, your messaging might not be the only factor you need to improve. Consider testing different images, color schemes, and calls to action as well.

Once you have goals, questions, and assumptions identified, your initial use cases are set. These will change over time as your business evolves, but for now you have initial targets to build toward.

Infrastructure

With your use cases thought out, your next big consideration will be your infrastructure for performing analytics. You will need to store and analyze data and have the governance policies in place to maintain data quality. Creating a solid infrastructure is the core of building analytic capacity.

Data Storage

As you accumulate data on your team, you’ll need a secure place to store it. The primary options for data storage are on a local server, or in cloud storage.

A local server is essentially a separate computer that performs a specific task. In this case, you want a server for data storage, and probably to run your database applications.

A local server will give you complete control over configuration and security. However, it is also likely to have higher startup and maintenance costs. You’ll also need to hire someone to provide IT services. Additionally, you may be required to purchase additional server storage in order to scale the operation.

Cloud storage such as Amazon Web Services (AWS), Google Cloud Platform, and Microsoft Azure provide managed solutions for data storage. These platforms are regularly backed-up, highly scalable, and have flexible pay-as-you-go pricing. The drawbacks to cloud storage are ongoing operational costs and a dependence on internet accessibility. Data security can also be an issue unless you purchase high tiered service.

Data Analysis

One of your biggest decisions when starting or expanding your analytics team is which analytical software they will use. There are many options you can choose from, and there is a lot of overlap across tools. Perhaps the largest divide in software categories is between paid software and open-source (i.e., free) software.

Paid software packages like SAS, Stata, and SPSS have been in use for many decades. As such, their interfaces and processes are very robust. These tools often come with many built-in options and user support that will help analysts manage your data and complete analyses. Additionally, programs like SAS are well-suited to large data sets (i.e., think 100s of GB) and fast processing.

Unfortunately, paid analytic software tends to be very expensive. This expense for paid programs, however, does not generally translate into greater functionality.

In contrast, open-source programs like Python and R are free to use, even for commercial purposes. Both programs have robust user communities that are always developing new packages and functionality, many of which are also free.

Additionally, Python is a full-blown programming language, capable of doing many things aside from analysis (e.g., application development). In contrast, R is a statistical computing language with high-end math libraries.

Open-source programs tend not to be as fast as some of their paid counterparts. However, you won’t notice a slowdown in speed unless you are trying to use very large data sets.

You should explore your industry and ask if there is a preferred software (e.g., SAS is preferred in healthcare). Otherwise, you might wish to consult with your analysts and outside experts to determine which program to use.

Data and Analytic Governance

If you are spending the resources to collect and analyze for strategic decision-making, treat it like an asset. Data governance is the set of policies and procedures used to control how data is collected, stored, analyzed, and disposed of. Data governance also defines which staff can access specific data, and to maintain compliance with regulatory standards.

Your goal in developing strong data governance policies and procedures is to facilitate accurate, useful, and secure data. You want those data to be accessible by the right staff at the right times for strategic decision-making. Solid data governance provides a defense against improper data access and use, and greater democratization of data.

Analytic governance is a similar concept to data governance. Your analytic governance policies and procedures focus on how analytic resources are allocated and managed across projects. Additionally, analytic governance sets expectations for the lifecycle of data analysis projects.

Your goal in developing strong analytic governance policies and procedures is to manage analytic resources to benefit the organization. Additionally strong analytic governance creates greater consistency in development, execution, and reporting from analytic projects.

Data Quality

As you are developing your data infrastructure, always keep data quality in the back of your mind. After all, we’ve all heard the old saying: garbage in, garbage out. The last thing you want to do is develop your infrastructure and find out your data stinks.

When you think about data quality, focus on how you are collecting your data and ensuring its accuracy. The five characteristics of high-quality data are accuracy, consistency, timeliness, comprehensiveness, and uniqueness. If you develop processes to maintain these characteristics in your data, your analytics will have a promising foundation.

To help troubleshoot data quality issues, you should focus on standardizing your data capture processes. You want to ensure that each type of data you have is collected in the same way from one record to the next. For example, make sure that phone numbers are captured with 10 digits and are consistently formatted with parentheses or dashes.

Another important process you should consider for high data quality is regular data review. Set a schedule to check how much data you have. Review the data to make sure all your fields are consistently captured. Double check there aren’t any duplicate entries, and that all your fields contain only values you expect to see.

Methodology

At this point in the process, you’ll need to begin thinking about your methodologies. Specifically, which analytic strategies will you use for each question you developed earlier in the process? How will you document your analyses? And what procedures will your team use to ensure that the analysis and deliverables created are all accurate? Your decisions here will form a major block in the foundation as you build analytic capacity. But never fear, these can be changed later as needed.

Analytic Strategies

Each analytic question you ask will require at least one analytic strategy to provide an answer. Typically, this is where your statistical analyses come into play. You will need to work with your analytic team to determine which analyses they will pursue.

At this stage, you may feel nervous about making decisions if you are not a strong analyst. However, identifying useful analytic strategies for your project is different from understanding the underlying math or being able to program the analysis.

You only need to understand what an analytic strategy does, when to use it, and its basic assumptions. The analysts you hired on your team can handle the math and the programming.

Documentation

All projects require documentation of what happened and why for future reference. Analytic projects are no different. In fact, complex analytics can be even more demanding of good documentation.

There are 5 critical analytic documents your team should create for each analytic project. These include:

• Project Process Document (PPD): tracking all files used in the project, staff responsibilities, and process for creating the final deliverable.
• Analytic Plan: defines exactly what the analysts will do during the analysis. This is the plan that the analysts develop. Any deviations from the plan should be noted in the PPD.
• Technical Specifications for Measures: this document defines the calculations used to create any metrics.
• Timelines: a calendar of project milestones, deadlines, and dependencies to keep everything on track and identify bottlenecks.
• Start-Up and Close-Out Checklists: these documents provide your team with easy-to-follow steps that ensure your analytic governance standards are followed.

Quality Assurance

For many professionals, the idea of quality assurance stops with the person who inspected their last online clothing order. Data analysis, however, requires keen attention to detail both during the analysis and when compiling deliverables.

You want your analysts to develop a plan for verifying that their results are correct. You should not assume that they will do this on their own, because analytic validation is often not taught in school. There are many ways analytic validation can occur, which generally trade-off time and expense for accuracy.

You also need to develop a plan for verifying that your final deliverable is compiled accurately. Regardless of whether you create a report, slide deck, interactive dashboard, or social media post, your credibility hinges on accuracy.

Ask someone to review each deliverable to verify tables and figures are in the right locations. Also ask someone who did not write the report to read it with an eye for whether the text ties back to the data correctly. As a third level of review, ask someone who did not work on the report to review it for clarity and proper explanation.

Team Development

If you’ve made it to this point and don’t have anyone on your staff in an analytic capacity, then it’s time to start hiring. You want to make sure that you hire staff at the correct level for what you need. Additionally, you’ll want to invest in training capacity to help retain good analysts and have them grow with your team.

Staffing

Staffing an analytics team is not always easy at the beginning. You need to balance costs with capabilities; and analytics can be expensive. That’s why teams tend to use one of two strategies to build analytic capacity.

The first strategy you could use is to hire the team lead first. Then you provide a budget and set parameters around what the analytics team needs to do. By doing this, you effectively delegate team development to someone else quickly.

An alternative strategy is to hire junior analysts first to set everything your infrastructure and documentation. With guidance from existing leadership or a fractional Chief Data & Analysis Officer, you can accomplish these two tasks easily.

Then give your analysts small projects to develop that will familiarize them with your data and performance monitoring results. As your analytic needs grow, you can hire in staff with greater expertise or provide training to bridge the gap.

The choice of how you develop your team will generally be a trade-off between higher costs for expertise up front with quicker results, or lower costs up front with a slower timeline to results. Let your specific circumstances and needs guide you here.

Training

One of the reasons why good analysts leave jobs is to pursue more advanced opportunities elsewhere. If you are able, providing analysts with training to develop their skills will improve both their satisfaction and loyalty. The result will be a greater return on building analytic capacity.

Begin by listing out the types of analyses that you want them to be able to complete. Then put that list into order with the simplest analyses first. Have your analysts identify the level of analysis for which they are not comfortable creating and executing a plan. The remaining items on your analytic wish list will tell you what your analysts need training on, and in what order.

You might run across analysts who want you to use whatever the latest, shiniest, sexiest analytic methods are (e.g., AI, deep learning, neural networks, etc.). Be sure to confer with an expert before agreeing to this approach or providing any training toward it. Many simpler analyses can provide information that is equally useful and a fraction of the cost.

Still, if you can afford it, training your analysts beyond your current needs can still reap benefits. This is especially true if they identify new avenues for analytics with high return on investment.

Data-Driven Culture

With the fundamentals developed and staff in place, you’ll want to begin cultivating a data-driven culture in the team. When you develop a data-driven culture, you will create a feedback look to reinforce your efforts to build analytic capacity. Because culture takes time to develop and refine, you should expect that this domain will be completed last in the process. Although, you and your team will be working on creating the culture throughout the process.

The two top elements you want to foster in your data-driven culture are curiosity and good communication.

Curiosity

One of the primary attributes I look for in an analyst is curiosity. They should be inherently interested in understanding how things work and why. A curious analyst can learn a new industry and business relatively quickly. Additionally, curious analysts tend to be more thorough than those lacking curiosity.

Still, none of this means that you cannot foster curiosity in your staff. You want to actively encourage your staff to ask questions, and prompt them to do so regularly. You also want to encourage your staff to use the data they have available to find answers to the questions they ask. By doing this, you will build an expectation for asking and answering question.

Keeping your staff engaged with their work will also help foster curiosity, or at the very least allow it to flourish. Encourage your staff to seek mastery over the work they are doing. Give them opportunities to take ownership over elements of their work that are suited to their skills and knowledge. Allow them to provide feedback and recommendations, taking these comments into consideration in decision-making.

Communication

A second key element of a data-driven culture is effective communication. You need your analysts and leadership to be able to communicate clearly and efficiently. If they fail to communicate well, expect substantial delays and rework on projects.

To foster effective communication, you want to make sure everyone is on the same page about your analytic project. This means that your analysts have a thorough understanding of the business problem you want to solve. Similarly, you want leadership to understand enough about the analytics to be conversant about the project with the analysts.

In addition to the foundations of substantive knowledge, analysts and leadership must develop special translation skills. Specifically, each group must be able to translate their specialized knowledge into terms the other group will understand.

For leadership, this means parsing the business problem into elements that help the analysts design and execute the project. For analysts, this means translating technical results into plain language explanations answering specific questions.

Conclusion

Your interest in building analytic capacity demonstrates your commitment to growth and success. Using data to identify problems, and develop solutions is perhaps the quickest way you can make operational improvements.

Developing analytic capacity on your team may seem like a daunting task. However, the process is actually far easier than many leaders realize.

The 14 elements described in this article provide a broad view of key factors that you will need to consider. You can plan out many of these elements in a few short meetings with your team.

If you are not experienced in building analytic capacity, you may need to seek outside assistance. Find an expert to guide you through navigating the decisions and standing up your infrastructure. You might even want to hire an analytic hiring consultant to help you review resumes and interview candidates.

If you are interested in additional information, or want to discuss your specific circumstances, schedule a call with me at F1 Analytics. I’m happy to help discuss what might work best for you; and I promise there will be no high-pressure sales.