How to use Voice of Customer at a PLG Company: Insights from Apollo.io’s CPO
Understanding the Voice of Customer (VOC) is vital in the fast-paced world of Product-Led Growth.
Recently, we had the chance to sit down with Chief Product Officer, Abhishek Viswanathan from one of the most successful PLG companies, Apollo.io. During the Roundtable: Connect Customer Feedback to Revenue Impact Abishek shared how Apollo.io set up a successful VOC program and how they have use it to 9x revenue and reduce support requests by 40%.
Read on to get the the highlights from our Q&A Session that will help you understand how to get started with a Voice of Customer initiative.
Question 1: What advice do you have for product leaders starting a voice of customer program?
Abishek: Focus on a core problem and tie the VOC program to specific business outcomes. Design the program around these outcomes for more tremendous success. You need to have a core problem you're trying to solve. You should not set up a VOC program just for the sake of a VOC program.
At Apollo.io, we wanted to focus on retention, and one of the input metrics for retention is self-serviceability or reduction in friction. And we tried to debunk the input metric into multiple themes, like what is the reason why there is friction in adoption? If you have a business outcome tied to a set of input metrics that a VOC program can support, then you should design the VOC program around it. If you have a generic VoC program just getting customer insights and not tied to a particular business outcome, it is less likely to succeed.
Question 2: How do you get executive buy-in for a VOC program?
Abishek: Use data to demonstrate the program's potential impact on critical metrics like gross dollar retention and cost reduction. Design the program for specific outcomes to strengthen the case for implementation.
I'm a math person, so I recommend setting up a funnel. Suppose your human inquiry rate is high. In that case, you can look at the cost of service, and you can also attribute that every time there is an interaction that a customer has with a support channel, it reduces their gross dollar retention. So if you take the combination of both of those and if you can estimate the impact of, if you reduce the inputs to friction through a VOC program, you understand what the inputs of friction are. Here is how you can impact overall gross dollar retention at a minimum. It reduces the cost and opportunity cost as a contact center, which you can use for things that are much better for expanding the business.
The best way is to convince yourself that it's worth it by estimating the impact. Once you do that, you design the program for that specific outcome. At Apollo, we knew the human inquiry rate correlates to retention, so we said if we reduced that by 40%, this would impact retention in the long term. Getting buy-in was more of a tactic after that.
Question 3: Which team set up the central data repository for the program, and do you have any tactical suggestions?
Abishek: Our BizOps team played a central role, ensuring that the program had all necessary data and reporting in one place. Regular meetings and clear accountability helped drive the program's success. Yeah, consider this like a program. The setup of the data is one part.
We set up a centralized program to have all of the data and the reporting in one place. So every product team could access that. And then we also had governance around it. Every team had signed up for a goal, so we would meet monthly to understand what we see in the data. What are the top themes? What is it that we have fixed in the last month? How have our trends for human inquiry rate gone in the previous month? And we are driving accountability through that single forum where we have all the PMS.
Our very talented BizOps team runs the program and the cross-functional coordination. The BizOps team runs the show because it's rare that the Product team singularly owns fixes. Messaging, marketing, website, knowledge base articles, support interactions, and fixes needed across the company exist - having BizOps support that centrally was phenomenally impactful for us.
Question 4: What was the most challenging part of evolving your VOC process?
Abishek: The most challenging part is the program setup and governance. I wish somebody had told me, here are your input metrics and the outcome metric. So we had to sit down and brainstorm, and two folks from our BizOps team, Nathan and LindLi, were extremely helpful here. What would the right outcome metric be? For everybody, there will be a different outcome metric that you might have to go and create. That was the most challenging part—the most surprising thing. The most surprising thing was not very surprising. Every team had precisely three or four issues they had to fix. That significantly dropped the human inquiry rate. It was maniacally similar to three or four issues. It was a short list. So before, we had that long list of everything customers were complaining about; it was so overwhelmingly long that teams were almost scared of touching it.
After identifying the Pareto, figuring out the Human Inquiry Rate impact, and seeing themes from the Enterprise cohort, every team had three to four issues. The number of issues is low. You can focus on fewer issues and get a ton of impact if you connect revenue data with themes across multiple channels and tie it to a particular outcome, like the Human Inquiry Rate.
Question 5: How do you weigh and prioritize different sources of feedback?
Abishek: We prioritize feedback based on total license value and customer profiles, ensuring we focus on the most relevant and impactful feedback for our target customer segment. We still look at it a little bit more in-depth with, is this the right customer profile we should even be serving?
Our product is such that we have specific profiles of customers who might pay a lot of money, but they need to start using the product for the core jobs to be done that we are helping customers go after. We use ARR quite a bit to look at prioritizing, but we also look at the customer profile to see if this is the tier of customers we want to serve. Because you can go and fix the root cost, but the underlying problem might be much more significant. The product is not suited for that customer profile. That is saying that this doesn't work for them. So it's a combination of those.
Want to read a full recap of the roundtable? Check out this blog post, How Apollo.io reduced support rates by 40% and maximized revenue with customer feedback.