girish mathrubootham, freshworks
Girish Mathrubootham, Founder and CEO of Freshworks

The one area of business where data is separating the winners from the losers at the fastest clip is the race to understand customer behavior. Across sectors, data intelligence is increasingly seen as the most potent weapon in the battle for customer acquisition and retention. Little wonder that from likes and shares on social media to browsing habits on shopping apps to service tickets for complaint resolution, companies are furiously mining all kinds of customer data.

This gladiatorial battle for data raises several questions: How do companies prepare themselves for this brave new world, given working with data needs specialized talent and skilled hands are in short supply? How indeed do they attract the best to work with them? How do they balance data with human intuition? 

Shanmuga (Shyam) Anandaraman, freshworks
Shyam Anandaraman, Director, Product Management, Machine Learning and Data Sciences at Freshworks

For Girish Mathrubootham, founder and CEO of Chennai-based Freshworks (formerly Freshdesk), a company that is disrupting business software using the software-as-a-service (SaaS) model, these are fundamental questions. If Freshworks’ growth is anything to go by, Mathrubootham is answering them with aplomb. Since its launch six years ago, Freshworks has notched up over 100,000 customers around the world who use its business software to improve efficiency, enhance service quality, and shrink costs. The company now has offices in San Francisco, London, Sydney, and Berlin, making it one of India’s most successful and fastest-growing cross-border startups. (Full disclosure: SocialCops is a Freshworks customer.)

With scorching growth like this comes the thorny question of decentralizing the organization’s knowledge – and that’s where building systems to enable the free flow of all mission-critical data becomes vital. Over exhaustive conversations with us, Mathrubootham and Shanmuga (Shyam) Anandaraman, Director, Product Management, Machine Learning and Data Sciences, Freshworks, give us a glimpse into how the company is making this happen and the hurdles along the way.


Edited excerpts:

Girish, what do you believe is the biggest reason businesses should adopt data-driven decision making, and what is the right mindset to have when you go on this journey?

Girish Mathrubootham (GM): Well, businesses tend to make a lot of assumptions about their customers — who they are, what they want, what they need. It’s only when you have the right data that you know whether any of those assumptions are correct. [With regard to your question on the right mindset,] I believe in the importance of the human hand in customer engagement. Any business that tries to use, say, machine learning, to understand customers without human intervention has failed. In my experience, the assisted model [where technology supports human decision making] works best.

Shyam, Freshworks serves hundreds of thousands of customers across the world, solving their ever-changing CRM needs. It’s a complex web. Given this context and your role in the company, how have you seen the importance of data intelligence grow?

Shanmuga (Shyam) Anandaraman (SA): Let’s start at the beginning. Take a product manager like Girish, who has spent over a decade in the customer-support industry. He has acquired a huge amount of domain knowledge, extending to every nitty-gritty of the product. To understand this, think about a travel-booking website and the various buttons that you see there. Right down to the text on each of those buttons – “book flights” or “discover deals” – the placement of the buttons, or the design of dropdown menus, every element is decided based on nuanced, in-depth knowledge of how to deliver an effortless experience to the end customer.

Data reduces the dependence on domain knowledge. That’s the primary reason every company today is working to make data as transparent and easily accessible as possible within its decision-making teams.

In the initial days of Freshworks, Girish could pass on all this domain knowledge to the team and say, “Build the product like this.” He spoke the same language as the end customer, the team was smaller, and communicating what’s needed was much easier.

But now, the company is expanding. We have MBA grads working with us, for instance, who are smart people and can figure things out but do not have a customer-support background. We cannot keep relying on people with domain knowledge if we want to manage our growth. That’s where the data on how customers use our product every day becomes critical. Data reduces the dependence on domain knowledge. That’s the primary reason every company today is working to make data as transparent and easily accessible as possible within its decision-making teams.

That’s an interesting way of looking at it. The usual narrative around data is that it is a tool to design better solutions for your customers, but as you put it, data’s primary role is to empower your own people to do their jobs better. That’s why critical data has to flow freely within the organization…

SA: Yes, and it’s a complex process. In the early days of any company, the critical data and knowhow is with a few individuals because it takes resources and time to build and scale the systems to ensure that the data flows freely. Also, the priority is to chase growth and revenues. But once the company matures and it doesn’t have to worry about revenues, the focus shifts to retaining the customers and acquiring new customers. That’s when data becomes important. It’s a critical transition point in the company’s life. It has to transition the knowledge that’s owned by a few subject-matter experts to a wider group of people…

This can be a difficult transition. How do you make it easier?

SA: Right. Holding on to data is a way for people to build a moat around themselves and preserve their importance in the organization. You have to do two things to change that attitude. First, you have to show them that the more they share, the more everyone – including them – grows. Second, you design systems that don’t have people dependencies.

Nobody should have to ask for the data they need from someone else. Mature companies like Google have created data lakes, where all the data resides in an anonymized and aggregated form.

There’s a team whose job is to make it possible for any user to query any data, even if they don’t have, say, an engineering background.

We too have a dedicated team whose job is to unlock the potential of all the data we have for our internal users first. If we can use data to drive marketing optimization, improve ad placements, or SEO, the revenue needle moves, the impact is clear, and everyone is happy.

Could you share an example of how you are using data to improve your customer experience?

SA: Today, organizations manually sift through tickets (in CRM parlance, a ticket is a service request raised by a customer) and decide whether they are low-, medium-, or high-priority. This is time consuming. Now we are building an app that can auto-prioritize tickets by looking at past data. Currently, it has 85%-90% accuracy for some our customers, and we are looking to pilot it soon with them

Building all this requires a workforce skilled in working with data. Given the skyrocketing demand for data professionals, why should they consider joining the CRM industry?

SA: Data professionals can have an outsize impact on a CRM company’s growth since data is absolutely central to it. The two pillars of the CRM business are sales and support. Both are heavily data-dependent. The day-to-day work of a support agent involves processing a lot of data on past customer issues and how they were resolved. Similarly, salespeople need to analyse a lot of data to identify, nurture, and convert leads.

Consider the process of scoring leads. A salesperson can log into his inbox, and the dozens of emails sitting there are already auto-sorted according to how hot the leads are. Imagine what that does to his efficiency.

Today, we are already achieving such improvements with a fairly high success rate just by using third-party, open-source tools and frameworks. Data scientists don’t need to be PhDs. We simply need smart engineers who have the aptitude to learn.

Girish, let’s switch focus. This is a question we like to ask every entrepreneur. As the leader of a fast-growing company, how do you maintain the balance between your instincts and cold, hard data?

GM: If you sit in on enough meetings, you know that if you torture any data enough it will tell you what you want to hear (smiles). I often talk about Jeff Bezos’s letter to Amazon shareholders, where he describes the culture of Day 1 that makes Amazon a unique company. That’s where most startups are — on Day 1 of their journey.

This is the time when you need high-velocity decisions and high-quality decisions, even though you may not have all the data, neither can you afford to wait for all the data. Yes, you have some data which shows you that for a certain problem, you have to find a new solution, a different way of attacking it. But from there on, you often depend on your gut to show you the way.

Debjani Ghosh, who recently stepped down as the chief of Intel in India and South Asia, told us that one of the big changes that data-driven decision making is bringing about in corporate cultures is that it is making bosses accept that they might be wrong. What is your take on this?

GM: I think accepting that one might be proven wrong from time to time doesn’t necessarily have to be a function of data. Leaders falter when they want to appear invulnerable, because that leads to insecure behavior. Of course, if eight out of 10 decisions you make are wrong, then you must admit there’s a serious problem with your leadership. But if eight decisions are right and only two are wrong, then it’s okay. You move on. Also, remember that a lot decisions can be reversed. So there’s no point feeling insecure.


rE


Photo credit: Freshworks and Shyam Anandaraman