From Complex to Simple: Q&A With Bedrock’s SVP of Engineering, Nav Bhachech
The Bedrock Analytics platform is simple to use and presents users with easily understood insights on the sales trends of thousands of CPG products. But behind the platform stands infinitely complex datasets that are made even more complicated by the variety and discordance of its sources.
Needless to say, the Bedrock engineering team deserves a lot of credit. Our team of software engineers and data experts work tirelessly behind the scenes to make the complex seem so simple. One of the people who helps this process run smoothly is our SVP of Engineering, Navdip Bhachech — a multidisciplinary leader who has built and managed teams for companies like Amazon, Microsoft, and many more.
To give you a better idea of what goes into Bedrock, we asked Nav to share a few details about the work that goes into making syndicated and retail portal scan data easy-to-understand for analysts and non-analysts alike.
Tell us about your role at Bedrock and your work on the platform.
I am responsible for the engineering team at Bedrock, which includes functions such as program management, testing, operations, and of course, development. The platform is a cloud-native data and API platform, and we’re evolving it to use more flexible data pipelines, a pluggable AI, and more automation than before.
What is Bedrock’s core philosophy when it comes to data analysis?
The guiding principle in our approach is to make data analysis as simple as possible for the user. As part of that, we also aim to make it easy for our users to share the analyses and insights with others, whether by creating internal reports or custom sales presentations for their retail partners. Bedrock organizes CPG data into distinctive storylines designed to build a narrative that helps explain not only what is happening, but how and why it is happening. Our system has built-in intelligence that pre-calculates interesting analyses for the user, so they don’t have to wonder what the charts and graphs are trying to tell them. There is also a lot of effort put towards making the generated analyses easily shareable in different ways.
What essential skills do you draw on in your day-to-day job?
One of the most crucial skills in my job is flexibility and adaptability. As with many tech startups, my team and I are often asked to change the hats we wear day-to-day, and sometimes hour-to-hour. We are a small but mighty team, and while we do have focus areas amongst the team, we still trend towards a multi-stack approach in which any member of the team might be asked to chip in on any number of areas.
I’d say another essential skill is the ability to step back and think strategically about the issue at hand. When we are innovating so quickly and trying to push new features and functionality into the marketplace, it’s easy to get wrapped up in tactical efforts. But we try to carefully consider every decision we make with the big picture in mind. At the end of the day, if something doesn’t contribute to our mission of turning complex CPG data into simplified selling stories, then we have to abandon it.
What makes Bedrock unique as a CPG platform compared to other analytics tools on the market?
We focus on enabling the CPG business person to derive the maximum value from their data. Whether they’re in sales, business intelligence, category management, marketing, product development, or the executive ranks, we turn complex CPG data into actionable insights to help them understand the trends affecting their own products as well as those of their competitors, and the entire industry as well. Our platform is among the most advanced analytics tools available for any industry, but it is custom-built to address the nuances and intricacies of the CPG market specifically.
From an engineering perspective, what are the obstacles to making complex data easily digestible, and how do you overcome them?
The engineering challenges come from keeping the analyses accessible and adapting them to the customer’s view of the world. So we may need to view the data with their custom category hierarchy in mind, which could change a lot about the data. Keeping the data flexible and making it useful in different visualizations and storylines is a hard problem! In order to help alleviate some of that pressure, we keep the data in different formats and levels of readiness by prefetching and pre-calculating what we think the user might need.
One important Bedrock feature is data harmonization. What considerations go into creating pipelines from diverse sources and unifying their metrics?
There are lots of things to look out for there. Each data set may have specific assumptions about time periods and cycles — for example, their week might start on a Sunday or a Monday, or their quarterly cycle might be 12 or 13 weeks. The formulas used for the same calculation may be subtly different, or the formatting may be different — yielding different results. Or how they define dimensions such as markets might be different, making it hard to compare across data sets and sources. So we’re always thinking about those.
What are some of the trends currently unfolding in data analytics and SaaS spheres? How is Bedrock reacting to them?
One analytics trend is that we’ve gone from calling it BI to calling it AI. One letter better, I suppose. But seriously, we’ve been able to apply better technologies in order to automate some of the insight generation we provide, and we moved from advanced slicing and dicing to deliver on the promise of mining data for insights.
Another trend is SaaS apps — even the serious ones (a.k.a. Enterprise) are merging with consumer apps in terms of customer expectations and features. So SaaS apps have to look at things that they might have previously ignored in order to meet customer expectations.
What are your predictions for the CPG analytics space within the next three to five years? What will change, and what will stay the same?
Data will become more embedded into CPG brand lifecycles. We’re seeing more brands adopt data even before they have major distribution, and we think that the smart brands will begin to leverage data from the very start. This approach will enable them to plan their launch, get insights to grow, reduce discontinuation risk, and maximize customer lifetime values.
Data services will also evolve to be easier to use. They will blend with much more diverse datasets and provide lower-priced entry points (like the Nielsen/Bedrock free trial!). This democratized data will continue to create opportunities for players who can monetize the value of their data assets in this landscape.
What’s next for Bedrock from a software engineering perspective? Any exciting new features or updates planned?
We have a pretty exciting roadmap, but are moving away from being feature-driven to being scenario-driven. It’s part of the process of growing up, I suppose. And the evolving world of CPG data we just talked about will drive our roadmap. We’ll also spend a lot of time making incremental improvements to existing features so that existing customers continue to derive more value from using the software.