Pleo's proactive data strategy for customer success

Tarush Aggarwal (00:00):
Hi, everyone. Welcome to another episode of People of Data, where we highlight the wonderful people in data. These are our data leaders from various industries, geographies, and disciplines. Discover how companies are using data, their achievements, and the challenges they face. Today on the show, we have Sriram Sampath. Sriram is an experienced Head of Data, leading data at PLEO. PLEO, as we all know, is a massive fintech, B2B unicorn. But what I'm most impressed by about Sriram are his extracurricular activities. Sriram has recently completed a thousand-kilometer cycling trip across India, during which he focused on building six mini libraries for children. He also spends his time volunteering for math and science education with the Lil Einstein program at Bhoomi, India's largest volunteering organization. Welcome to the show, Sriram.

Sriram Sampath (00:55):
Hi Tarush, happy to be here. I've also seen some of the past guests on this show, so I'm honored to be here. Thank you.

Tarush Aggarwal (01:02):
I think the honor is all ours. We'd love to get started. I know we have a lot to talk about, so let's begin with what PLEO does. We all know PLEO as a B2B unicorn, but fundamentally, what does PLEO do and how do you guys make money?

Sriram Sampath (01:16):
Yeah, sure. So Pleo is a centralized expense solution for modern teams. It helps your employees spend the money they need for their work while giving bookkeepers full control and visibility of the expenses made. In a way, it caters to two major personas: the spender and the bookkeeper. Typically, companies come to us with traditional challenges, and we make it simpler. You have shared cards used within a company, missing receipts causing issues for bookkeepers, and cases where employees chase employers for refunds and reimbursements. Overall, it's always a laborious task to put together expense management reports for the bookkeeper. What we offer is one of Europe's leading B2B fintech solutions, making us the go-to for all things expense management.

Tarush Aggarwal (02:19):
How does data play a role at Pleo, and how does your team fit into what the business fundamentally wants to do?

Sriram Sampath (02:28):
Like any other business, especially a fintech, data plays a central role in our success and growth strategy. I can highlight three main use cases for data at Pleo:

Product Integration: Within the product, users can see a centralized overview of all expenses in real-time—where they're spending, what they're spending on, and who's spending. This makes a bookkeeper's task easier. We also automatically fetch receipts for recent payments and help categorize expenses in real-time.

Customer Insights: Data helps us understand our customers' needs and spending behaviors, especially in the current economic climate where customers may reduce or change their spending patterns. We use data to adapt and align with their needs, ensuring we offer the right products to the right customers at the right time through the right channels. This involves analyzing product engagement, adoption rates, qualitative research from user researchers, and spend patterns.

Risk Management: Data is fundamental in identifying risky customers and predicting fraudulent transactions, which helps protect both our customers and our company from reputational and financial risks. Additionally, it allows us to support our best customers by offering products like credit and overdrafts.

Tarush Aggarwal (05:19):
That makes a ton of sense. Just to recap, you spoke about three areas: one external-facing within the product, the second internal involving segmentation and personalization, and the third focusing on risk management to de-risk the business. That's awesome. Everyone talks about their business being data-driven, but we'd like to zoom into that a level deeper. What is really the culture at Pleo, and as a data leader, how do you fit in? How do you map into that? Does that make sense?

Sriram Sampath (05:55):
Yeah, in terms of culture, I think we mirror the product we sell. Our product is based on trusting employees and being transparent, both as employees and employers. That's the fundamental cultural aspect of Pleo—trust and collaboration to simplify processes for ourselves and our customers. Our culture revolves around trust and accountability. That's how I would sum it up in one sentence.

Tarush Aggarwal (06:29):
And how do you and your team fit into this culture? Obviously, we want to come from the data world to make data-driven decisions. How receptive is leadership to this approach? And if someone were to come work for Pleo.io, what should they know about the culture they're walking into, specifically within the data team, given the organizational climate you've described?

Sriram Sampath (07:06):
Absolutely. The culture revolves around buy-in from leadership; without that, you can have all the data you want, but it won't translate into action. We are a 45-member organization within Pleo, spread across four competencies: Data Engineers, Analytics Engineers, Data Scientists, and Data Analysts. Specifically, analytics engineering is an emerging sub-competence. Our leaders understand the value of building trusted and high-quality foundational data models so they can trust the data in their outcomes. There's been significant investment in this area to add more analytics engineers to our team. We have immense talent density, which is something I'm proud of and a reason I love working at Pleo. We push each other in the right way to achieve the unified goal of taking our customers to their next best experience. That's the most important part of working at Pleo.

Tarush Aggarwal (08:24):
I love that you mentioned being broken down into these four teams. You have about 45 people. What's the company's overall footprint? How many employees does Pleo have in total?

Sriram Sampath (08:37):
PLEO currently has 850 employees, with 45 of us working in the Data and Analytics organization.

Tarush Aggarwal (08:45):
Incredible. The data team represents about 5% of the company. Typically, we see about 20% of companies are in technology, and within technology, about 10-20% of the organization is in data. So, looking at about 4% of the org, you guys are somewhat larger than that, which is a good indicator that data is a first-class citizen within the organization. What does your tech stack look like?

Sriram Sampath (09:15):
Absolutely. Starting from data collection and ingestion, we use Fivetran and Segment. Our data warehouse is Google BigQuery. For data transformation and pipelines, we use dbt (data build tools) Core and Paradigm for lifecycle management. For data cataloging and governance, we use CastorDoc and Sensors for reverse ETL use cases. Elementary is used for data monitoring. In terms of data visualization and BI, we utilize Looker and Amplitude. So, that's a summary of our key tech stack components.

Tarush Aggarwal (10:27):
Amazing. All of the usual suspects. I love it. Are there any areas you want to invest in? Obviously, AI is on everyone's mind, and you mentioned one of your four teams is focused on AI. How does AI fit into your stack? Are you leveraging GCP, or are you looking at some dedicated external tools?

Sriram Sampath (11:04):
It's a mix of different approaches. We have an AI enablement team that sits outside the Data and Analytics organization. Their main focus is to identify use cases, both within the product and other areas, where we can leverage AI to achieve our goals in the next few years. We're consciously exploring how to integrate AI effectively.

Tarush Aggarwal (11:31):
Got it. What is one achievement that might not be obvious, which you're super proud of? In other words, what is one use case you want potential data engineers and data professionals to watch? What is that one unconventional use case you've cracked that you're particularly proud of?

Sriram Sampath (11:52):
Yeah, I wouldn't say it's unconventional, but one of the use cases we're particularly proud of at Pleo is an engine we've built that enables our commercial organization to reach out to the right customer at the right time with the right offer through the right channel. This engine taps into both qualitative and quantitative insights, making it a unique selling proposition of our data product. It doesn't rely solely on quantitative data, which is the traditional approach. Instead, it incorporates what our customers are saying on the ground. Our sales and commercial customer success managers gather this feedback and tie it back to the engine. This allows us to use data analytics, data science models, customer behavior patterns, and quantitative inputs from the commercial organization to create a well-rounded, holistic engine. This helps our frontline teams have meaningful conversations and proactively reach out to customers before issues escalate or before customers realize they can benefit more from a product feature they're not fully utilizing.

Tarush Aggarwal (13:14):
Yeah, sure.

Sriram Sampath (13:33):
This engine has enabled over 300 organizations internally. It taps into over 20 signals from different parts of the business, such as product engagement, adoption, and growth potential of customers. By integrating these signals, we provide the right enablement opportunities for our customers.

Tarush Aggarwal (14:12):
That makes a ton of sense. This is the sort of Customer 360 problem, right? It sounds like you've tapped into not only the obvious aspects like visit frequency and platform interactions but also the qualitative pieces, which often come down to sentiment. It's super exciting to see a true decision engine built on top of Customer 360. I think that's something more and more organizations need to tap into. That's very inspiring to see.

Sriram Sampath (14:51):
Yeah, just to add to that, this engine reminds us that we shouldn't work in silos. It's a collaborative engine where we tap into both data and context. The context from our customers and customer success teams is critical, especially at this hyper-growth stage. Typically, people on the frontline and those in data work in very siloed environments, each feeling they have all the necessary information. I believe that combining data with context solves most critical challenges that a customer or company faces.

Tarush Aggarwal (15:37):
That's so powerful. I think any consumer-facing company aims for a comprehensive Customer 360 because it's inherently multi-departmental. Every interaction a customer has with you feeds back into this engine, demonstrating the level of collaboration needed at an organizational level to execute a project like this.

What are the challenges you face today? What's something that keeps you up at night, or what are you excited about solving next?

Sriram Sampath (16:26):
There are many challenges, but one key issue is setting up our customers for success during the early stages of their onboarding or initial interactions with Pleo. How quickly can we gather these signals to extend the right level of service and provide an optimal onboarding experience? This ensures they're set up for success for the rest of their journey as a customer.

The challenge is twofold:

Customer Confidence: We want our customers to feel confident that Pleo can solve their use cases early on.

Human Capital Efficiency: Our human capital is at a premium, so we need to ensure we're investing our resources with the right customers at the right time.

We're striving to not only set our customers up for success but also intelligently leverage our internal human capital. We've made significant strides in the past few months, but there's still work to be done.

Tarush Aggarwal (17:55):
I understand. Mapping back to your Customer 360, initially, when you don't have vast data points to get the decision engine up and running, what do you do? Lookalike segmentation probably plays a role. How do you approach solving this problem? You mentioned making significant strides in the past few months. From a tactical level, what has that approach been?

Sriram Sampath (18:35):
I'm a big fan of keeping things simple. Using Pareto analysis as an example—where 80% of challenges can be solved with 20% of actions—we've adopted a similar approach. We start with the right baselines, understand which levers are critical, and build on that to ensure sustainability and scalability without causing operational confusion. We aim to start small to provide enough evidence to back the success of our initiatives while identifying the right signals early on. This prevents us from taking complex approaches without knowing what's truly making an impact. We've come a long way in identifying key signals, and now it's a matter of building on those baselines. We're making progress, but there's still work to be done.

Tarush Aggarwal (19:49):
Very interesting. Sriram, thank you so much for being on the show today.

Sriram Sampath (19:53):
Thank you, Tarush. It was a pleasure talking to you.