VP of Product at Apollo
Krishan Patel shares how Apollo moved from a sales-led motion to product-led. He discusses why the decision was made to start a PLG motion, how they ensured their free offering did not cannibalize sales revenue, the time and team it took to build their self-serve product, and what it takes to create true retention.
Krishan Patel is the VP of Product at Apollo.io. He started in sales as an enterprise SDR at Oracle and moved up to AE before joining Apollo and getting involved with Product. It was at Apollo that Krishan actually led the initiative of starting a product-led motion at Apollo when it was very sales-led.
Why did Apollo go from a sales-led motion to experimenting with product-led?
There’s the why and then there’s the how.
The why is when we took this software to market about six to seven years ago, we found that although we had some success in landing these midmarket enterprise companies, like Lyft or Mercer, before our series A, we got a ton of interest from the SMB or like SME businesses that wanted a one-stop shop to build their go-to-market strategy on.
We realized that we had extreme product market fit within this segment because these SME companies didn't have the time or the budget or the Ops teams to be able to cobble together a bunch of different tools that don't talk to each other and try to learn how to use them and set them up correctly to actually start building their go-to-market strategies.
When we look at these other tools, like Outreaches or ZoomInfos of the world, for example, they target these companies. They set up their pricing where if you can't afford a $9,000 annual contract, you just cannot work with these companies. We also had out pricing set up this way. We realized that there was an extreme underserved market in the SMB where we had extreme product market fit with SMB in a way that's better than our competition. At the time, ZoomInfo and Outreach might have been better for the mid-market when mid-market companies wanted a point solution that could be really good at solving one problem and they weren't getting the full benefit from our edge, which is being an all-in-one solution where our using our products in conjunction makes each product stronger. The SMB companies really, really wanted that.
Although we had product market fit, we didn't have product market channel fit. Although these companies really liked us, they didn't want to talk to a sales rep and they didn't want to pay $9,000. They didn't want to be locked into an annual contract. It took a lot of internal debates and conservative experimentation just to prove that this could be a viable strategy for us. We knew that if we could execute this fit, effectively we'd be opening ourselves up to dominating in this SMB market and capturing this whole area that nobody else was going after. That was a lot of the initial hypothesis that went into us deciding whether or not we should even experiment with self-service. We knew that if we did want to go after SMB with our current team, the unit economics just did not add up to be able to serve these SMB customers at such low contract values.
How did you avoid cannibalization when adding a free motion?
(Krishan) We built our free product to balance internal politics. However, if I were to do things differently, I think how conservative we were might have been a bit of a mistake. We probably could have grown faster, more quickly, if we were a little bit less conservative.
We started by building a self-serve product that had no monetization ability just to see whether all these people that sales was DQing could be put into some type of product where they're able to retain and continue being a user of Apollo instead of sending them to some of our competitors in the SMB, but without cannibalizing a lot of our existing business.
(Francis) One of the risks that comes to mind when hearing this is how do you make sure that you're actually building a self-serve version of the product for an ICP group? The risk of offering something to people that are being DQ’d is that potentially then you start optimizing for a category of companies that are where you don't have product market fit. How did you make sure not to fall into that trap?
(Krishan) When we looked at companies that we wanted to be users of a platform, there were like two types of people that we DQ’d. One would be founders from SaaS startups that just generally didn't want to buy into an annual contract, but they knew how to use modern tools or they were going to use somewhat similar tools that might be more affordable. They really wanted to use Apollo because it was more advanced than a lot of the other solutions that you can get for free or for a cheaper price.
Then there's like this second bucket of people who were DQ’d where they're doing things like wanting to reach out to the plumbers in their own town or something. It's just a very different, much more niche use case. When people fit in the former, that’s when we took a lot of the data and qualitative interviews and research. We definitely made sure that we're only talking to the people that we felt were part of the core ICP of where we wanted to go in the future.
There were times, especially now, where we got tons of users that are not necessarily in our ICP. People find new use cases for Apollo all the time. We get students looking for jobs trying to find hiring managers to reach out to them and so on. When we take a look at our quantitative metrics, we're able to slice them by things like persona, company size, industry, and so on. We really just look at company size and persona though. When we're doing qualitative research, we're very specific about who we talk to.
What were some of the milestones working on the self-serve initiative?
(Krishan) The first thing we focused on was activation retention. We wanted to make sure if we build a self-serviceable product, that people can get to the value without ever talking to a human. We wanted them to form a habit around unlocking value from Apollo so that we became a bit more ingrained in their day to day jobs. At that time we still had a shared team. It was me plus some of the shared engineers, usually around two to three engineers working on this.
Once we got to a decent place, when it came to NPS of this self-serve platform, activation, retention of the platform, we started testing monetization. And again, we were super conservative. We initially tested monetization by putting up a landing page. When people clicked on it, they got an email from myself, where I'd ask for their credit card information. I’d process it manually in Stripe and they become a customer.
In the first month we hit 120k ARR, which was as much of our as our top producing sales reps. That’s when we decided that there's definitely some opportunity here, because this is much more scalable than hiring a bunch of world class sales reps. That's when we ended up shifting a lot more resources towards this. We were focusing o unifying our mini self-serve product with our existing product because or NPS and our user retention was better for our self-serve product than for our existing product.
We realized that when we're making improvements on the growth side, if we're going to be building the product to be really good and much more sticky, we might as well have that be part of the same product that our reps are selling. It actually would make sense from the monetization perspective to figure out a pricing tier where we're going all in on PLG.
After a lot of time spent on some platform work to unify the platforms, we started focusing on acquisition. Then the team scaled back down to maybe two engineers and myself to build this product-led SEO product used to generate tons of visits. We went from 6,000 website visits per month to 2 million after we launched this product-led SEO product. Once we got to that point, that's when we decided to fully build a separate growth team and growth squad. We knew that we always wanted to be testing hypotheses really quickly on this growth team, and we knew that we had every part of the funnel complete at that point. We had solid activation, decent retention, although looking back on it now, it was terrible compared to where we are now. At the time, we had some monetization and good acquisition at least.
We felt that now that we had some type of flywheel, we could start putting a team together to consistently and constantly try to iterate and figure out where the highest leverage gaps that we can influence to really bolster this flywheel.
At that point, it was probably nine to twelve months in before we actually built that first full on growth team.
(Francis) How long did the process take from getting started to seeing the first fairly consistent credit card swipes?
(Krishan) We started unifying our platform maybe four to six months in or so into the process. Once we unified our platforms after that, we were still very conservative with our self-serve unverified pricing. I think that was one of the mistakes we made along the way. We kind of treated it as making sure rep driven revenue stays constant or grows, but anything we do on self serve is in addition to that. So at that point we had conservative pricing. We were still, you know, closing like a hundred grand per month, sometimes two hundred grand a month for a few months.
Around months six to nine we started realizing that rep driven revenue was not increasing very quickly. It also was not being cannibalized too much from the self-serve product, but we realized that we had a lot more user demand for the self-serve product. We had tons of free users and there was a lot of demand to be able to move on to a tier that was between the current self-serve offerings and the free offerings. We started doing more research around pricing and talked to a bunch of users, and so on to figure out what optimal pricing looked like if we're actually trying to optimize for total revenue rather than optimizing for rep-driven revenue first and self-serve revenue afterwards.
Once we conducted the research there, we actually ended up splitting our pricing tiers into having like a $50/month plan, a $100/month plan, and then custom self-serve enterprise plans. We knew what are those features that enterprise users need that we can put on customs? What are those features that those middle ground users might need that those lower paying users might not need? Then how do we figure out what goes on custom versus pro versus basic and so on.
Once we launched that, that's when our self-serve started massively taking off and finally started to bring in more revenue than sales. And I think since then we've gotten more and more bold with our self-serve revenue, where we started making really disruptive bets, such as providing unlimited email credits on a plan, which nobody in our space did.
And I think part of the reason why nobody in our space did it is because it's a really scary thing to do in the data space. If you use a data credit, if I find someone's email address, I might not have a reason to go back to your platform. So you can only do that if your user retention is really good and you have continuous value that you can deliver to your users.
I think at that point we actually did have that extreme product market fit, where people are going to retain, even if they use all their data credits, even if they find their entire ICP, because they still use us to engage with their prospects. That's when self-serve really took off. We went from like 5 million revenue to 40 million revenue. Now PLG definitely comes first in our business.
How do you address long term retention with the product?
It's a big miss in the B2B SaaS world when people try to address retention by locking people in annual contracts or multi-year contracts that provide no value to the user. That’s faking retention in a way. The fact that we moved to month to month pricing forced us to figure out to consistently provide value for our users. We were able to like 5-6 times our user retention by focusing on that problem.
When we talked to advisors for PLG from other companies, they always mentioned that retention is notoriously hard to move. It felt like something that we might just not be able to do. We studied a lot of the best practices from Brian Balfour at Reforge, or Andrew Chen, and other great people in the space who built frameworks around how to improve retention.
Building retention is definitely something that is possible, but you have to have the culture within your company to think about retention as how to improve value for our users, rather than how to make users keep paying us.
How did compensation plans for sales reps change so that they didn’t feel like the product cannibalized their potential revenue and not hitting quota?
(Krishan) The comp plans actually didn't change too much at the rep level. We did have to talk through what gets counted by sales versus what gets counted by the product team. We made it where that anytime sales interacts with a customer and then they convert, they get some credit so that they don't feel so bad about Product stealing their leads or anything, but we'll take some effort on their side.
The big shift that we saw on sales is that instead of sales trying to reach out to people to hard sell them, instead we have them focus on is how to you talk to some of our users and help them get the value that they want to get out of our free plans, or help them get the value they need to get out of the product so that naturally they'll convert. Then naturally the rep gets the revenue and the commission associated with that conversion.
It was a little bit different on the leadership side. There was a change for myself at one point where I used to just be comped on self-serve revenue, then I was comped on total revenue. After that, there was just much more interest in how we work in unison to do what's best for our customers. A lot of people at the end of the day still do want to talk to a sales team. It's helpful for us on the product side to have sales reps and humans that can actually go in there and help customers in a way that we can't. They can relay that feedback back to the product team so we can figure out how to eliminate that question from ever happening again. And now we are much, much more aligned.
Right now Sales leadership is not comped on total revenue. Back then we had a sales leader, his name is Ramsey Al-Ramahi. Great guy. He was unlike what you'd expect from someone who's a sales leader while you're going through a shift from sales to PLG. He had a lot of faith in our team because every time we ended up making one of these significant pricing changes, we realized that there was a huge influx in inbound and a huge influx of product-led leads. So it actually ended up not hurting the sales team, but the transition felt scary. The people at the rep level were scared the most any time we wanted to make a change. But Ramsey definitely did seem to see the vision. He would try to rally his team along.
What we did do on the comp perspective as well is was if there was some change that we made where it crushed sales ability to hit their quota, then we gave then quarter leave for that quarter because we knew that it was something that we did directly to harm their ability to hit quota.
now it's a little bit different where our sales leaders are still not comped on like total revenue and it's like rep driven revenue. But me personally, I'm comped on things like How many meetings are our SDRs setting? So I have all sorts of things tying into my metrics, but I have to be very aligned with sales and supporting sales, and I like to be
True, best in class product-led growth means that you have really good self-service ability, but you also have a really good product-led assist motion. Similar to like Figma who just got acquired for 50x multiple. They had PLG figured out both on the self motion, but also on how they use their sales reps.