Less is more: Inside Personio’s powerful, lean data team

Tarush Aggarwal (00:00)

Welcome to another episode of People of Data, where we get to highlight the wonderful people of data. These are our data leaders from various industries, geographies, and discipline. Discover how companies are using data, their achievements, and the challenges they face. Today on the show, we have is a seasoned data leader with a wealth of experience in driving data strategy and analytics at top tech companies. Currently serving as the VP of Data at Personio.

‍

Adrien (00:08)

Okay.

‍

Tarush Aggarwal (00:26)

He oversees data initiatives that power smarter decision-making. Previously, Adrien was the senior director of data and analytics at Outreach. He's also had leadership roles at Zendesk and JustAnswer. Welcome to the show, Adrien.

‍

Adrien (00:39)

Thanks. Thanks for having me.

‍

Tarush Aggarwal (00:41)

Of course, are you ready to get into it?

‍

Adrien (00:43)

Let's go.

‍

Tarush Aggarwal (00:44)

what does personia do?

‍

Adrien (00:48)

Cool. Personio is a SaaS company. leading HR software, so specialized for small and medium sized businesses. it's a HR platform. a software that makes a HR platform for those customers. Today, we are a little bit more than 1,800 employees. think it's...

‍

You can see online we have a current valuation of 8.5 billion, which is making one of the most exciting startups Europe. mission of the company is to unlock the power of people for organizations around the world and echoing what I just said earlier with the goal of making small and mid-sized organizations more successful. So to do those, we offer...

‍

I would say we have four tonnets that we follow to offer our customers. One is to offer customizable solution, which they can adapt to their needs. So needs to be convenient and easy to set up solutions. So lot of flexibility in our software. And the last piece is centralized information all and etc. software.

‍

HR. So that's a little bit the core of our product and we're currently in a phase where we're trying to expand and propose new products such as payroll and etc. cover the full suite of HR responsibilities through Personnel Software.

‍

Tarush Aggarwal (02:23)

Awesome. And does data play a role in helping Presonia with what they do?

‍

Adrien (02:28)

So, Personio is a data-driven company. they use data to one, I would say there's two obvious pieces, the internal and the external. There's internally to, and we're going to talk in shambles into details about it, which is like, how are we going to use data and helping to have a, I would say, a strategic edge across competitors and to know customers better.

‍

And then there's a piece about the customer facing product and how do we use data to have reporting with all of our customers. So data are core, I would say, to everything that's happening at Personio from an internet decision making to facing product.

‍

Tarush Aggarwal (03:20)

Awesome. And what's the culture of the company like? You're obviously one of Europe's biggest unicorns by valuation, very, very exciting, lot of attention on the company, especially in this market. What's the culture like working internally?

‍

Adrien (03:33)

It's amazing. a lot of really driven people have ambition that want to do the right thing to make the company successful. So everyone is really accessible. Everyone has a focus on trying to do what's best for the company collaborate and work together to try to

‍

to achieve it. There is, have as a company, think, really ambitious goal and ambitious leaders, and everyone is running back and trying to deliver. also at a stage where not really a startup anymore, but we still have a lot of a startup mindset of trying new things, trying to launch new product, being reactive.

‍

et cetera. it's a great place to be from a people level really interesting, I would say, place to be from a job work aspect.

‍

Tarush Aggarwal (04:35)

And how does the culture of the data team map into this? Very often data has got this very unique macro view lens of the company, but at the same time also wanting to be a little bit more systematic and sort of forward thinking. So I've always found that data teams have this very interesting blend of culture. was wondering, what's that like at your group?

‍

Adrien (04:52)

Yup.

‍

We

‍

and we have to let me to explain a little bit. think it makes probably sense that I explain a bit of background of data, the department at Personio So data used to be, would say decentralized a little bit within the company. like the executive team and leadership team decided to centralize data in one department. And where I came to.

‍

to lead the department. we have the culture actually within data to want to push the bar, raise the bar and try to continue to progress. are all in all a young department because we're probably like 18 months old and we were trying to push things forward and

‍

drive impact. answer your question on specific culture, we do and I think every data team needs to have that specific culture. Why? I think it's a really unique role and it's a really interesting role. There's not many or responsibility within the company where people need to be technical.

‍

but also really business-minded. a great data team, in my mind, is a team that can drive impact that works with all of tools and all the data to drive that impact. And from a culture and people standpoint, to drive that impact, we need people that have those technical abilities, but that also have that business connection, business sensitivity, understanding, and have the ability to do the junction between the two.

‍

So from a cultural standpoint, it creates a unique culture because it's people that needs to be really honestly smart, driven, and et cetera, and that create a really unique and cool, I think, environment.

‍

Tarush Aggarwal (06:47)

Yeah, that makes a ton of sense. the topic of being technical, what does your data footprint currently look like?

‍

Adrien (06:55)

Yeah. we are currently, I think, between 40 and 50 people. So the way we organized, often internally use the image of a restaurant. We are a burger restaurant. So we have the data platform engineering team.

‍

the one that the groceries. So basically they manage our pipeline and manage our platform. Then we have the analytics engineering team. The team that pick up vegetables and clean them within the restaurant. So, know, chop them, cook them. in reality, they manage DBT, they manage layers, they manage some self-serve initiatives and to ensure that people can self-serve across the company.

‍

Then we have different analytics teams that we call embedded. we have product analytics, customer analytics, and PWA analytics. And finally, we have two other teams. One data science that's also doing AI machine learning. And we have, which I think is fairly unique and great personally, have a data strategy and ops, which is a team around internal operation.

‍

around cross-functional program management we are trying to carry within the team the junction between the rest of the teams also act as a scrum master. So that's from a people's standpoint. Then from a technical standpoint. Pipeline, mean, a lot of data is stored in AWS. For pipeline, use Airflow, we use Fivetran Then...

‍

For data warehousing, use Snowflake. visualization, we use Tableau. Then we have different tools. We use Atlan for data catalog. we use for data observability. I think that's all tech stack.

‍

Tarush Aggarwal (08:51)

I think you mentioned DBT earlier, but I'm assuming DBT...

‍

Adrien (08:54)

Yeah, we

‍

do have DBT, sorry. exactly. Before, we have DBT as well. Who is DBT, of course.

‍

Tarush Aggarwal (08:58)

Also,

‍

very, very sort of all of the usual suspects over there, which is exciting to see. mentioned the entire data team is about 40 people. How big is the company?

‍

Adrien (09:07)

Around 1800 people. So we are between 4 and 50.

‍

Tarush Aggarwal (09:11)

Very typically you see about 20 % of a company is about technology. So, you know, add about 18 people at the ballpark about, you know, 350, 400, and typically about 20 % of your technology team is about data. So if I had to guess, you know, typically, you know, 60 to 80 would be

‍

Adrien (09:34)

Listen.

‍

Tarush Aggarwal (09:34)

about

‍

that. it sounds like it's a lean team, know, using best in class tools to be highly productive. So super exciting. I just love, you know, getting this.

‍

Adrien (09:46)

Yeah,

‍

that's correct. We have a little bit. could add, if you take data overall, you could add probably like five or 10 people that are probably part of a product. yeah, it's roughly. we are lean and we trying to stay that way. I want to ensure that we...

‍

we value and that we have well-oiled machine before trying to the engine.

‍

Tarush Aggarwal (10:12)

Yeah.

‍

No, I think the world is changing. I wouldn't be surprised if that 4 % number becomes 2 % in the next few years. I recently wrote a post on this, but think we're moving towards a team of... I think we're moving towards rise of leaner, more end-to-end, more productive teams. So that makes a downside.

‍

Adrien (10:35)

I agree.

‍

I'm not sure. But I agree with change of roles. I think we'll see probably a shift with AI, more automation, and et cetera, of role changes. I don't know if pure headcount within data is going to change drastically. But I foresee, for example, the old school BI analyst reporting, role to actually

‍

become less and less prominent and moving towards probably analytics engineering to prepare and like, know, to have that reporting layer that's usable for AI where the self-service becomes more and more easy. And the move of the analyst to be really strong strategic partner and becoming more and more focus on the business. exactly that role of middlemen just producing.

‍

Tarush Aggarwal (11:12)

Yeah.

‍

Yeah.

‍

Adrien (11:25)

dashboards and et cetera to probably shrink in the future.

‍

Tarush Aggarwal (11:29)

Yeah, no, that's I think a lot of people have been speaking about this and I think it'll be very interesting to see how this evolves in the next 18 months. What's one achievement, as you mentioned, it's a relatively new org, been there for about 18 months. What's one achievement which you're very proud of, where data has had a big impact on the business?

‍

Adrien (11:51)

Yeah, I think there are a few actually. going to break it out because I actually see data as different, I always say building blocks and we don't have one responsibilities. I think first we have, which maybe against what I just said, but like the reporting responsibilities. We need to ensure that people across the organization

‍

can have access to report, can have access to data, hopefully in some sort of some sort of manner, but to ensure that they can make the right decision. And so that's part of work we did around like, you know, it was obviously existing already, like consolidating or landscape and showing that we have the right reports, that we have the, and that it's built in a scalable manner, which I think is something that's really important too. Then there's another, and so that's...

‍

I don't see that as a huge achievement, but it's an important one. Then have a part where we are involved in the day-to-day operations. quickly, I forgot in our tool stack, we also have what we call a reverse ETL, Hightouch So we are actually involved on rewriting data in all our system across the organization. Like that we have one source of truth.

‍

And we built what I think is actually a fairly strong achievement from a personal standpoint, the CDP, which is our customer data platform. So everything our data warehouse, we actually built a full CDP where we can have all the different attributes to each customer and create that source of truth for every system within the company where we can, that.

‍

is aligned and uses the same thing. So we write to HubSpot, to Salesforce, and etc. But like that, the customer definition and the customer attributes are centralized and managed in one place. I think it's important of the data team that often are not always agreed, actually, in different data teams I actually work with or manage or discuss with.

‍

Some data teams don't want to be part of operational process. I think for the future of a data department, are going to need to be part of those, thanks to also the technical skills we can bring. Then, I go to the next piece of what did we achieve, there's a lot around analytics. So we have like we build like different scoring model, like...

‍

cross-sell model, like the potential of each customer to actually cross-sell to a new product that's actually used in different parts of the company to try to be more efficient. one similar to that would be what we built what we call an adoption score, which is like trying to predict with that customer at that risk of either churning.

‍

is our reduced engagement and we can develop a playbook our customer support team to try to reach out and understand happening with those customers. would say one thing that I'm really proud of actually that we also delivered. we actually started doing data accaten. So we everyone for like a few days.

‍

And that's one of the projects that came out of the Akaton, but it was really well received. So it actually made it as a data product, only a One of the things that Personio is really focusing on, it's customer love. So we want to receive all the customer feedback much as possible and have a way to ingest them and take action on it.

‍

That's something that's a big focus for the company. One of the challenges of that was actually we had many different sources for customer love. you can have Zendesk of your tickets, you can have your different surveys, your C-STAT, your NPS survey. lot of different places where we could hear Salesforce, the text from Rep and etc. Where we could hear customer feedback.

‍

And aggregating it within for an analyst is doable, but it was also reducing our ability for self-service because people could look into like Zendesk and etc. we weren't able to, we were trying to find a solution to join everything everyone within the company could basically interact. So what we, what the team did,

‍

During the hackathon they built actually a chat GPT of interaction where we ingest all of customer feedback in one place. And we have right now we call our VOC, which is our voice of customer, employees can go and ask questions about different customer behavior, what does customer feel, and et cetera. And summarizing everything, taking all the different data sources we have in one place.

‍

giving the feedback to whoever ask it. So you can have a product manager that goes and say, hey, what are the three main feature requested for, I don't know, time management? And you're going to have the GPT that will answer to you based on all different sources across the company.

‍

Tarush Aggarwal (16:57)

I love that voice of customer that is such a awesome use case and I can totally see how it's going to be so popular with your product managers. Is that using, that using ChatGpt or do you have a different type of LLM layer?

‍

Adrien (17:06)

Yep.

‍

No,

‍

we use Cortex Snowflake

‍

Tarush Aggarwal (17:14)

Amazing beautiful.

‍

I think you know, Snowflake Cortex is very exciting. This is the first time on the show we have we have a customer using Cortex.

‍

Adrien (17:22)

Yes, we are one of the first, think, based on feedback I heard from Snowflake

‍

Tarush Aggarwal (17:26)

Awesome. What is one challenge which you've had which you're currently solving or you previously solved?

‍

Adrien (17:34)

would say, sorry, I'm gonna, two challenges. think two challenges we're facing. Us being young, think have been successful, but still it's ongoing trying to do different things at once, which is like set up our infrastructure to be efficient, ensuring that we have the right reporting in place. And at the same time,

‍

already, I would say, being ahead in our maturity curve as a data department of providing insight, providing scores, impacting the outcome of the business, influencing customer-facing data, et cetera. So one of the challenges is to ensure we can cover all of this and we are really good at all of this. cannot...

‍

We cannot just be okay at all, otherwise we are going to fail. So we need to always raise the bar everything we are doing and ensuring that we bring the right value and we bring the right product The second piece is around scalability and it echoes a lot of the first, which is like, to your point earlier, we want the team to be lean, but we don't want to, you know...

‍

shortcut anything or don't do everything. So it's about like trying to find the right balance being scalable in our infrastructure, ensuring that we prioritize right and work on what matters the most, what will have the most impact and et cetera. So I think those are two main challenges we facing today data. I think it's, you know, it's standard.

‍

based on all maturity curve. But something then.

‍

Tarush Aggarwal (19:10)

Awesome, Adrien, thank you so much for being on the show.

‍

Adrien (19:12)

Thanks a lot for having me. Good luck to the show.

‍