TABLE OF CONTENTS
Disclaimer: This Blog is a work of fiction 😉. Any resemblance to actual analysts, workplaces is purely coincidental 😊. The only thing that is definitely real is the pain-point that analysts feel on a daily basis.
Hey there, I’m Andy 👋
I stepped into the exciting world of eCommerce as a business analyst around three years ago. Let me take you through my journey and share my experiences.
Entering the company, I was eager to dive into petabytes of data and empower the organization to make informed, data-driven decisions that fuel growth.
While the first part held true, my work wasn’t quite the revolution I’d hoped for. Curious to know why? Bear with me a moment.
Getting started and quickly getting drowned in repetitive questions
In order to learn the ropes of the business, I began by focusing on understanding the key business metrics and their relationships with each other.
This understanding was relevant to answer pertinent business questions like “Why is my basket size falling and how can I fix it?" that category managers ask literally everyday.
Let’s take a deeper look at the approach I took to answer this particular question. I would start with “basket size” — an important metric for eCommerce profitability. To figure out the drop in basket size, I would analyze it across multiple dimensions like product categories, customer groups, cities, and app versions, to figure out where the anomaly is. But this was only the beginning of my quest.
Once I figured out the dimensions where there were potential anomalies, I would decompose “basket size” into its child metrics, which in this case were “items per cart” and “average selling price.” At a broad level, the whole idea is to keep drilling down until I reach root-causes that businesses can act upon directly.
I also realized that for any good data-driven org, these key metric monitoring and root-cause analysis needs to happen pretty much every day.
So yes, I found myself immersed in peta-bytes of data daily, but I was answering the same repetitive questions, that while extremely essential for business, were not really helping me learn much.
Unforeseen challenges of automating via dashboards
To automate this process, I thought of setting up a dashboard that business teams can refer to by themselves, thereby unlocking me.
I started with the vision to build a single dashboard that the category managers can refer to “one dashboard to rule them all,” but soon realized that in order to go from the top-level problem to the root-causes, traversing across different dimensions and child metrics, I needed to create a dashboard tree.
This created a new challenge for me to upskill all the category managers to navigate this dashboard tree whenever they had a question. Initially they gave it a shot, but eventually came back to me to figure out the “Why” of their questions. I don’t blame them because navigating multiple dashboards was an unnecessary complication amidst their already demanding tasks.
Despite my best efforts to build dashboards, I found myself as the sole user. Eventually, I gave up and went back to using SQL, even though it is time-consuming and exhausting.
Importance of a shared organizational context to problem solving
Another layer of complexity was when new business folks joined, there was an onboarding overhead that got created. A lot of times when this onboarding was not done diligently, these new folks were not completely aware of the existing metrics and frameworks and ended up creating new metric definitions and frameworks in the process.
This led to different versions of the same metrics floating around, causing confusion and errors. We could see this even in management meetings where validity of the metrics became a discussion point instead of how to act on them.
As a result, analysts like myself now also had to go around in circles, comparing different versions and fixing them.
Bottlenecking the process
Now, as the go-to person for business insights, bottlenecks emerge when I am unavailable or occupied with other tasks. I am frequently pulled away from strategic tasks that I want to focus on to address urgent business requests.
One possible solution to address this issue is to hire more analysts. However, it takes time to train new analysts and bring them up to speed on my context, which further strains my already-limited bandwidth. Not to mention, it creates a huge cost-center from the business perspective.
The need …. Abstract the mundane and focus on what really matters
Don't get me wrong, the eCommerce space is brimming with tonnes of high impact problem statements that I can work on, if only I had more bandwidth on my hands.
I just wish there were someone who can abstract away these repetitive tasks, so that I can focus on the cool stuff that I signed up for.
Is there a better approach to do this?
Are you someone who resonates with this story? In GobbleCube, we have just the right companion for you. GobbleCube is an LLM based guided analytics platform that can help you productize your business problem solving frameworks.
What this means is that next time business folks can simply ask their questions in plain English from GobbleCube. GobbleCube will understand the context and parameters of the question, identify the relevant framework to use in this context, and quickly run the analysis to provide the root-causes.
So now analytics teams can focus on more impactful tasks like defining new hypotheses, designing experiments and analyzing their performance, to unlock new growth avenues while GobbleCube takes care of all that mundane stuff.
Also, as an organization, GobbleCube lets you create a shared problem solving context, so that you can focus on taking actions instead of debating about which metric definition is correct.
If this sounds interesting, then give us a shout. We’d be happy to chat!
