Perspectives From The Data Wilderness

Are You Frustrated With The Analytics Process?

P1000194 reduced.jpg
 

Are you managing an analytics team or project but you’re new to analytics?

Do you wonder why some analytics tasks take so long?

Is it difficult to pin down project timelines?

Let’s take a closer look at that analytics process to get a better understanding.

The Manager’s View

P1000258 reduced.jpg
 

It’s not uncommon for companies to have an experienced manager who is not an analytics expert oversee an analytics team or manage an analytics intensive project. This can be a frustrating position to be in for someone who is not familiar with the analytics process.

It can be notoriously difficult to get accurate time estimates for analytics projects, which makes it seem like projects take too long. It can be hard to determine what’s really going on, whether efforts have gone off course, or if you need to step in and how to redirect or refocus your team when necessary.

Let’s look at things from the perspective of the analyst and explore ways to work more productively with your team.

The World Of The Analyst

P1000281 - reduced.jpg
 

As a manager it can be hard to understand what’s going on at the ground level and your analysts may find it difficult to explain. A better understanding of the analytics process may help.

Analytics is a multi-step / multi-Day Process

Even a basic analysis requires several steps usually spread over multiple days to get familiar with the data and the problem, conduct an initial data evaluation and exploration, and complete a first draft of the analysis. Any requried modifications or updates require feedback and additional time and steps.

Analytics is Also an iterative process

Analytics can be more art than science. Large or complex problems don’t have one obvious approach, method, or solution. There’s a lot of trial and error, which makes it hard to estimate how long it will take. Sometimes a team stumbles upon a great solution immediately while other times it takes multiple attempts.

Intense focus is Required

It is not uncommon for analysts to be working on multiple projects at the same time although this is usually not a good idea. Analysis requires getting deep into the details while maintaining a sense of the big picture. Mentally keeping track of all of the details makes it very tricky to work on more than one analytics project at a time. Once an analyst stops work on a project it takes time to re-immerse in the details and continue working.

Analytics Is a Team Sport

It can be very frustrating and unnecessarily time consuming to work alone as an analyst. It’s easy to get stuck on a particular method, a bug in a software program, or get too far into the details and lose sight of the big picture. Without a team, or at least another analyst to help, it takes a lot longer to get unstuck.

Data Quality Can Be unpredictable

When navigating corporate data it is not uncommon to encounter data quality issues like inconsistent, ambiguous, and missing data. Data cleaning can take a significant, and uncertain, amount of time. Even when data has been used before and has been cleaned, it may not have been used in the same way requiring additional data preparation.

The Challenges of Writing Code

Many analytics efforts require some level of computer programming such as SQL queries for extracting data from databases, and R / Python programs to process and analyze data or create a repeatable analysis. This remains true even as tools have become more advanced and require less coding. Similar to analytics, computer programming is also multi-step, iterative, and requires intense focus.

The Pain of Debugging

Debugging is a big part of coding and involves running code multiple times to find and fix problems. It's hard to estimate when you'll discover a bug or how long it will take to fix. Sometimes solving one problem creates a new problem adding to the time uncertainty.

How Can You Work More Productively With Your team?

P1000300 - reduced.jpg
 

Whether your project is report development, report automation, a new analysis, or a new model, now that you have a better understanding of the challenges you’re in a better position to address them.

Know Where You’re Headed

Before any data is analyzed be sure that you and your team understand the problem (ensuring that you’re solving the correct problem), then determine what sort of solution is expected and in what format.

What does the deliverable look like?

A set of numbers, an Excel Spreadsheet, a dashboard on a preferred software platform, or a full text report? Creating a rough mock up of the final deliverable in whatever format it’s expected in will help get everyone on the same page. It will allow you to iterate quickly on the design and general content before moving ahead with more complicated analysis so you can show quick early progress.

One Time or Repeatable?

If you create something great that people love they’ll want to see it again. Most one time requests turn into requests for repeatable analysis or dashboards, which requires more work, forethought, documentation and a plan for future maintenance / support. Be prepared for this likelihood.

Working as a Team

Analytics is a team sport. You need a team of analysts and business experts working together, even if they don’t all report directly to you. If you only have one analyst consider partnering with another team with one or more analysts. It’s easy to lose sight of the big picture when you’re deep in the details, have your team work together to maintain the right balance.

Coding As A Team

If your team does a lot of coding it’s best for the team to decide on one language that everyone uses. Establish guidelines for adding comments and documenting code. Well documented code should be clear enough that someone with business knowledge but no coding experience can generally understand what it does. Finally, have your team review, run, and test each other’s code.

Break the work down

Limit multi-tasking in favor of more focus. Asking analysts to switch between projects means more time to re-immerse in each project later. However, there are natural break points in an analysis. Work with your team to figure out where the break points are so work can be paused or handed off to another analyst.

On Project Management Tools

Let your team decide how best to share work. There are many project management methods, but if they don’t work well for your team than they’re a waste of time and resources. Let your team choose the one that works best for them.

Ask For a Tour

If you’ve never written computer code but you manage projects that involve computer programming it’s a good idea to sit down with one of your analysts and have them walk you through what they do to get a better sense of what the process is like. You don’t have to learn to code or understand everything, just a glimpse into their world will help.

More Understanding = Less Frustration

P1000170 - reduced.jpg
 

The analytics process can be challenging for managers and analysts. A better understanding of the process and what it’s like on the ground level makes it easier to work more productively with your team and reduce the frustration for level for both you and your team.

Do you need advice or guidance in building and managing an analytics function? Could you benefit from data analytics coaching or training? Do you need to improve the analytics capabilities of your team? Contact me at stacey@arielanalytics.com to find out how Ariel Analytics can help.

About the Photos: Lucifer Falls, Buttermilk Falls, Taughannock Falls, Ithaca, NY
Thoughts: The finger lakes region in upstate New York has beautiful moderate hiking where you can see some spectacular waterfalls. It’s not exactly wilderness, so get ready to climb some stairs.