Breaking Down Information Silos with Atlan
The Active Metadata Pioneers series features Atlan customers who have recently completed a thorough evaluation of the Active Metadata Management market. Paying forward what you’ve learned to the next data leader is the true spirit of the Atlan community! So they’re here to share their hard-earned perspective on an evolving market, what makes up their modern data stack, innovative use cases for metadata, and more.
In this installment of the series, we meet Pranav Gandhi, Head of Data & Analytics at Signifyd, a leader in eCommerce Fraud Protection technology supporting thousands of stores in over 100 countries. Pranav shares how a company built on data science will use Atlan to break down information silos, driving fast, confident decision-making for technical and business users, alike.
This interview has been edited for brevity and clarity.
Could you tell us a bit about yourself, your background, and what drew you to Data & Analytics?
I lead Analytics Engineering and Data Analytics at Signifyd, and have been at the company for about four and a half years now.
I got started in Data & Analytics when I joined Jet.com, an eCommerce retailer that was acquired by Walmart. When we moved to Walmart, I pivoted into pricing analytics, which aligned with my background in Economics. It fascinated me to see how data could be used in so many ways and different functions.
Would you mind describing your data team?
Signifyd is unique in that we’re a Data Science company first. It’s our product, and isn’t a means to an end. We make money when we provide decisions. Our team is uniquely organized, and there are active conversations about operating as a data product team.
So, we have a Decision Science team, sitting in a different part of the organization but utilizing a lot of data to help make those decisions. Our data team is essentially part of our product organization, and we treat data as a first-class citizen within our organization, akin to a product.
My team is made up for Analytics Engineers, who are hands-on with data and creating models for others to use. Then there are Analysts, some of whom are centralized and help teams like Product, Marketing, Data Science, and Finance. We’ve already begun decentralizing some analytical functions in a hub-and-spoke sort of model, and they’re already reaching the scale where their coordination with our centralized Business Analysts and Analytics Engineers is working well.
Why search for an Active Metadata Management solution? What was missing?
The way our teams were initially set up was creating silos in how we managed our information. Root Cause Analysis could also add additional complexity for our data teams, even with simple asks. We’re also constantly testing and releasing new products, which means the way customers send us data changes frequently. The data team sits far on the “right” of all this, and some context was sometimes missing, so we would have to ask questions in Product and Engineering channels on Slack. That took time and put pressure on our analysts, especially those who work to make our customers successful.
If the customer isn’t being served in an optimum way, that can be a drag on their business. So, making sure people had access to the right information and understood it was paramount. We also realized that there were so many siloed ways of organizing data, that it was even harder to have a clear way to exchange information across them.
So, we started to look at centralized cataloging tools. We thought about Looker, because that was the primary place where our data landed, but found it was too “late” in the data workflow for that information to live. That’s when we started to consider Atlan.
When you were evaluating the market, what stood out to you? What was important?
In the Active Metadata Management market, I think there’s an identity crisis from a lot of vendors. Are you solving for technical users to understand their workflows better, or are you solving for business users who have no clue what these concepts are?
What was tough for us is that we wanted our choice to solve as many use cases as possible, because we want to be cost-efficient in order to scale in an optimized manner. We couldn’t afford to have a tool that only solves Data Engineering and Analysts’ pain points, while leaving the business users in their own silo when they’re the users who could benefit the most.
When we talked to different vendors during the evaluation, the biggest thing we learned was that if you aren’t solving for both personas, then you have to assume the business user isn’t going to enter the tool. With Atlan, there’s the Chrome Extension, so business users don’t have to worry about needing to sign into a new tool. With the other approaches, you can create personas, but usage isn’t going to be great all the way to the right.
For our more technical users, we knew they would use it. But we liked that Atlan had support for non-technical users, and it made it much easier for even a Data Analyst to do enrichment, as opposed to asking them to understand all the technical elements of how metadata is scraped before they could add value.
The place we landed in our evaluation is that Atlan had the product that sat most squarely in the middle between business users and technical users.
What do you intend on creating with Atlan? Do you have an idea of what use cases you’ll build, and the value you’ll drive?
We’ve started with collecting some business use cases and have a couple that are quite data-heavy where we’re creating things like customer health scores. These scores proactively help our customer success team understand information about our merchants. Getting people into one, central location where they can retrieve that information is going to help.
The way we’re thinking about this is that we’re not going to have a ton of users on Atlan right away. We’re going to roll it out by use case and we’re going to slowly enrich it, because it’s the sort of tool where if you move too quickly and things aren’t updated, then you’ve just created more technical debt in a different tool. At that point, you’re asking the question of whether bad data is better than no data. We don’t want that to be the case. So, we’re going to predominantly focus on business teams that come to the data team with a lot of questions.
Some teams have their own documentation, Confluence is used sparingly, and we’re a very Slack-heavy organization. We’re kicking tires right now to see what works internally, but we’re looking forward to having data contextualized and tagged on Slack via Atlan. I think it will be critical to get that set up correctly so users will see value quickly. We can also be more intelligent, and if we see that 20 users on Slack are asking the same questions about an asset, then we can prioritize documenting it.
Did we miss anything?
I would just say we’re looking forward to this journey. What I’m focusing on, especially in our organization where we value fiscal responsibility, is how we show value to the business and our internal stakeholders. You need buy-in to do something like this, and it requires change management. So, our team needs to make sure we’re getting the most out of Atlan, but also that both business and technical stakeholders are benefitting, too.
Photo by Bench Accounting on Unsplash
