In recent years, the trend of focusing on adoption and LTV have picked up serious momentum. At Unusual Ventures, we call this the Modern Go-To-Market (GTM). Historically, GTM consisted of isolated functions with their own metrics, in addition to supporting a sales team’s short term revenue goals. In the Modern GTM, overlapping roles and functions are focused on the customer’s journey and the long-term ROI they’re receiving from a product.
At the center of this trend is an extreme focus on product adoption, and beyond that, the level of engagement in the customer community. How much are people learning and teaching the product? Whether you call it a bottom-up motion or product-led growth (PLG), the focus has shifted to these metrics of adoption as an indicator of long term recurring revenue.
Unusual Partner, Sandhya Hegde, recently sat down with Arjun Devgan, Amplitude’s Global Head of Customer Success, to discuss organizing Customer Success for a Modern GTM Strategy, including metrics one should use as an indicator of long term recurring revenue.
Question from Sandhya: What did the post sales role look like at NetSuite in 2007? Did the function of customer success exist then as we know it now?
It was a very exciting time for the company. The professional services team at NetSuite was a very valuable function because we led customer implementations. We sold ERP, with a long convoluted process, to get customers up and running—many implementations can fail, so getting to the sale was only half the battle won. You need to get customers implemented and the way it works with ERP is, once you get them live, they will probably stay for a long time. Many implementations don't see the light of day. Customer success wasn't a thing per se in 2007. Post sales ran along the traditional model with services, support and account management.
Question from Sandhya: What did account management at NetSuite look like? Were there similarities with the modern Customer Success function or was it still more focused on account expansion and selling?
NetSuite was a unique beast. I have seen a lot of account management models from that era that have really great overlap with Customer Success. Our account management team was also a sales team exclusively focused on expansions. They were both effectively sales teams.
Question from Sandhya: Once the customer was implemented, the focus immediately shifted to expansion, arguably leaving the customer to their own devices in terms of driving adoption successfully?
Much of the onus of Customer Success was on the services organization—aligning with the economic buyer, which was usually the Chief Financial Officer. From making sure they understood what their organization's responsibilities were—to the implementation (which could take as long as three to six months.) Good CS teams needed to ensure that the product was actually being adopted properly, and that we're getting ahead of the renewal to mitigate risk. All of that was being run by the services leads like me.
Question from Sandhya: How have you seen the maintenance of relationships evolve over the past almost ten years now? When did you first hear of the function “Customer Success” whose goal is to focus on adoption?
Customer Success as a function grew out of the customers-for-life program in the 2010 timeframe at Salesforce. It really was one of the most important catalysts. Salesforce said, "We are a subscription software company, and we need to be more thoughtful and focused about how we drive lifetime value from our customers." That was the birth of Customer Success.
Immediately I was attracted to it. It put a new model in place, which said, "Let's put the customer at the center of the universe, and not your own expansion motion, and focus on how to make the customer so wildly successful that your revenues with a customer will grow rapidly year-over-year." And so it was about 11 years ago that I started to educate myself and evolve myself beyond a PS professional and into a broader CS leader.
Question from Sandhya: How has the role of CS changed from when it first started at Salesforce to now? Where was the explosion in data and automation?
Data is the key word. If you look back a decade ago, and regardless of what Customer Success was doing, people didn't really look at software consumption (adoption) on either the customer side or the vendor side. People were selling software and not monitoring whether it was being used, which was to their detriment but their customers weren't looking at it.
But what's happened over the last decade is massive SaaS proliferation. When I log in every morning, I have 15 different applications that provide me insights, and I'm probably going to really get into one or two. So when you look at the vendors of those technologies, they should all be concerned because they're all competing for my time and my attention. And that's true for all of their end users. And customers have an increased awareness of this. Today they think, "I should not have software sitting in my tech stack that I'm not getting ROI out of that is not being consumed."
So the confluence of that has really shifted the burden onto the vendors to say, "If I'm going to create software, I need to make sure it's adopted, and I need to equip the customer to adopt it successfully." And obviously, data is the first step towards understanding adoption. You need to understand what's happening within the application in order to drive that up. If people don't use your software, you're not going to get the renewal.
We did have lightweight usage tracking at NetSuite ten years ago, but it's become really a blessing for Customer Success teams now that it's become a more data driven business of looking at adoption.
Question from Sandhya: What are some examples at Amplitude of things customer success professionals are able to do today that you couldn’t do ten years ago?
Product data is life changing. So there's been three phases of it for CS teams. At first, there was no data. Then suddenly they had access to usage data. But what they did for a good seven years was look at monthly active users. It's immaterial. It's not a metric that matters. People are logging in, and you go to your economic buyer, they'll say, "So what?" If people are logging in, what are they doing? And now we've entered this new era powered by a bunch of systems, including Amplitude, where you can actually understand what they are doing. Are they doing meaningful things that matter, that drive ROI for them?
One great example of a customer success conversation we had at Amplitude is as follows: We charge customers based on event volume consumption or data consumption. And we saw some organizations that were blowing through their data, and you would normally say, "Oh, they should expand. They're consuming the data volume they purchased." If you actually go one level deeper to see whether that data is being consumed in a useful way, are those events being queried? It tells you that the customer actually should probably optimize their data volume before they purchase more software. They haven’t set it up properly.
So it's a really interesting insight in terms of how you could have had the wrong go-to-market motion to try and sell the customer more software. They might have churned instead of expanding because they now think product data is too expensive to track. Instead, you need to go one level deeper to make sure the customer is getting value.
Question from Sandhya: How do you see the relationship with sales evolving in this model and go-to-market function?
It's one of the hardest things for companies to solve for. I would say the bulk of the SaaS companies I've talked to have gone from an isolated and siloed customer engagement model to a much more integrated model. Thinking more holistically about sales, CS and all of the services/support and how they interact with the customer. Many people are trying to draw out roles and responsibilities of how an SC versus an AE interacts with the customer after they're a customer.
I think people struggle with providing clarity and often end up stepping on each other's toes. So you have multiple roles within the organization effectively doing the same thing or have gaps in the process of the customer experience. I think that's where customer journey mapping and clearly defining the resource model and expectations comes into play. It is a difficult thing to do and get right.
Question from Sandhya: How have you organized your whole function at Amplitude? How do you make sure that those early experiences growing customers are having are positive?
Yeah, it's a great question. I think it depends on what size of company you are. A smaller company (series A, B), there are a lot of generalist roles. You can onboard, train, and maybe manage the account end to end. We're certainly at a size with 1000s of customers where that needs to be differentiated. So we have five distinct teams: professional services team, Customer Success management team, education team, and a support team. We also have a programs team that focuses on building repeatable, scalable programs that are more technology enabled, that don't require a human-first motion. I think that's kind of appropriate for our size. When it comes to balancing and giving the customers what they need to be successful in a cost effective way, it all comes down to segmentation. You have to segment your customer base and create differentiated offerings, because you may not be able to fund all of the segments equally. Try and understand how much you can invest in each one of those segments and what ROI you expect out of that investment. That is probably one of the most important things you can do early. Obviously, smaller companies might only have two segments (mid-market and enterprise). Larger companies (like Salesforce) probably have dozens of segments and sub segments.
Question from Sandhya: How do you think about supporting a free plan?
Having spent a lot of time in the enterprise, I'm very passionate about free plans and the SMB segment. I feel like it gets ignored and it deserves a lot more attention. So let's talk about why the free plan is important in a company like Amplitude. It matters for a couple of reasons.
First, the free plan actually creates a lot of our pipeline for future deals - those customers become paying customers as they grow with a very low sales acquisition cost. Second, even if they don't buy—those people learn to use Amplitude and go to work at other companies. So there's this knowledge of Amplitude that spreads within the universe and expands the Ampliverse.
Free customers are not generating any revenue. So as a CS leader, I'm unable to deploy any people resources against that. But what you can do is create high quality resources for your paying (mid market and enterprise) customers that are reusable by your SMB and free plan customers.
Here’s a very specific example. We recently launched our Amplitude community. We are encouraging customers to go there. Not because we are trying to lower our support case volume, that's just a side benefit, but we think our customers can learn so much more when five, six or a hundred people are exchanging ideas on a question.
You can get some help through just a very narrow support interaction but how do you scale that? It's really about high leverage resources—making sure documentation, videos, webinars, and online communities are available. They actually benefit everybody in your broader community, not just the free customers.
Question from Sandhya: How do you resource mid market accounts (somewhere between 50K-100K) vs the potential million dollar accounts?
Let's start by just CS manager ratios. In the early days, most of your customers are mid-market/corporate and your product still needs a bit of work. Your CSM needs to lean into your customers to help them. You end up with a 30:1 ratio or 25:1 ratio. Probably not sustainable in that segment over a long period of time. So you try and figure out what components of the customer ask can be moved into a scalable, self-serve offering embedded into the product.
Examples might be CSMs doing a lot of 1:1 training that can then be packaged into e-learning or in app. Or perhaps your CSM is spending 15% of the time on questions that really should go to your support team.
One of the first things a startup should do when they design a full GTM organization is separate support from success—reactive from proactive, inbound and on-demand. That split is really helpful because it removes a bunch of noise from each team. Next, you're going to sub-segment your corporate/mid-market business. You’ll want to create a premium tier within your corporate segment for high growth customers who could become million dollar accounts.. We call these “named commercial" accounts. They have a much richer CSM and sales staffing model than some of the other customers because we can nurture and grow the account. Conversely, you might have accounts that are massive but very self sufficient—not growing and don't really want to talk to you because they're “saturated” accounts. You may be able to put a much lower touch model on them.
It really comes down to making sure you understand the customers because the mistake people make is putting the wrong customer in the wrong segment and they churn because they resourced it inappropriately. As you shift your CSM's time to focus on the right areas, you can start to get into the 50:1 and 100:1 staffing ratio.
One more thing that is really important to figure out fairly early in your journey is whether your product is going to require implementation services. CSMs who are required to do long implementations are not going to do anything else. They're not going to spend a lot of time driving adoption for renewals. So if your product requires a big implementation push, you may want to have an onboarding team and implementation team. You can charge a small fee and get that off the ground so you can have the CSMs focus on adoption and the multi-year journey of that customer rather than the first 90 days.
Question from Sandhya: What do you think are the right metrics for CS teams and some of the specialized metrics that separate some of these functions?
You should look at leading and lagging indicators. The leading indicators help you understand what is happening now, what is within my control now, how am I going to get ahead of it and are primarily tied to adoption and account health. So we have an adoption score in the system that is powered off Amplitude usage data. It looks at a few different indicators. It looks at consumption of event volume, it looks at active users as well as power users - people who are taking more meaningful actions within the application.
Some organizations tie those metrics to their core workflows. Sure these people are logging in, but are they doing the most meaningful action I need them to do within the application? Are they engaging in the way sticky customers and end users engage? That would form the basis for an adoption score that tells you whether adoption, keeping all other factors aside, is good. A more holistic customer health score would take into consideration things like community engagement. Are they attending events? Are they a reference for other customers? Is their relationship strong with the account team? The account health score is more holistic and can factor in more extraneous behavior. So those are three really important metrics—consumption, adoption and health.
Lagging indicators tie to revenue. None of the work matters if you can’t drive up gross dollar retention and net dollar retention. As a CS team, you have to be able to demonstrate that you're moving those numbers.
Question from Sandhya: What do you think are good benchmarks for net revenue retention, and what can you actually do as a CS leader to move the needle? What are the most important tools in the NRR toolkit?
It is top of mind for CEOs everywhere and I am grateful for that. It has put CS under the microscope in a good way. It's allowed us to really have the right conversations in the company around GDR and NDR. There used to be the famous 90/120 benchmark—you want to have 90 GDR, 120 NDR but people have really pushed the limits on that 120 nowadays. There are companies out there delivering 135 NDRs. And people love it. What that means is, if you didn't sell any new customers, you would still grow 35%. That is crazy. And we know the customer acquisition cost is so much lower for that expansion revenue. If you can grow your company 30% without setting new customers, imagine what your growth rate could be with the new bookings. The other interesting thing that has happened is, the higher you can go north of that 120 NDR, the more forgiving the 90 GDR barrier has become. But 85% gross dollar retention is just not considered good anymore, which is kind of shocking as it used to be okay.
In a pure enterprise business, you should be able to get GDR north of 95% and you will see the benchmarks on the NDR side, getting into the 140s and 150s.
SMB is a very interesting game, HubSpot has probably done the best job there. They were below 100% NDR for a long time but were so efficient in how they acquired (inbound marketing) and managed (self-serve) their customers that it worked well. They ran a very viable business and have NDR north of 100% now. Most companies do need to be north of 100% NDR, even at the bottom end of the market nowadays. It's tough when you're in the 90s and anything below that you don't have a business.
Question from Sandhya: What do you do as a CS leader, or as an executive in the company, when there is a retention problem in at least one segment?
Retention is not just a CS metric, it's a company metric. Sales is one of the biggest contributors to the NDR metric, and product development is one of the biggest contributors to the GDR metric. So, if I look at myself as a CEO or a CS leader, it's really about understanding the problem. Go look at the data and do some cohort analysis of segments. Look at usage trends and non usage trends, look at churn reasons, etc. I think the CS leader should be to leverage data and facilitate that conversation where product engineering, marketing, and sales can come together. You might go to engineering and say, "We need a much more stable product - the uptime expectations are outrageous in the industry we're in and the downtime we suffer is actually one of the leading contributors to churn.” To the product counterpart, you might say there's a major gap—we don't integrate into industry standard tools which is a problem and that's why people are churning and we're not able to expand.
You might need your marketing leader to realize that you are targeting some customers who tend to churn 30% more than everybody else. You have got to bring people together and drive that cross functional ownership of the retention number, not just own it yourself.
Question from Sandhya: Is there an example of where you saw a particular problem, recently at Amplitude or before, where you were able to change something in CS to address it?
Yeah, at one of my organizations, we had very bad SMB churn. Unfortunately, I inherited a decision where we were going to walk away from this SMB segment. Now, it may have been the right decision because it allowed us to focus our scarce resources on the enterprise.
At another company that I worked for, we had a similar problem and we went a little deeper and asked, "What is going on here? What is the reason that SMB churn is bad?" First, we had the wrong customers in the wrong segments from a resourcing POV. We had customers with a lot of potential and customers who had the need for resourcing in a segment that had very little CS resourcing. Second, we had poor incentive alignment. People who were overseeing those accounts didn't have the right incentives, from a compensation and an accountability standpoint, to actually do something about the problem. The primary thing we did was re-segment and restaff the model based on data and had more focus in terms of solutions. We saw five to six percentage points pick up just by cleaning that up and putting a really world class auto-renewal process for some of the low-touch customer segments. They just clicked a button and they were done. That segment performed phenomenally well, and became one of the best performing segments of the company. Somebody put thought into understanding what the problem was and actually deploying solutions to solve that issue.
Question from Sandhya: What would be your advice for someone who's adding their first CS leader to their team?
The one thing they should try and separate out as early as possible is to set up a separate support team. Don't mix support and success. The success hire can do a bunch of different things early on your hybrid resource. From a tool standpoint, you’ll need something that can help you pull your usage data early on even if it’s just spinning up a Google Sheet. You can start that simple if you only have a handful of customers. It’s important to learn how those customers are using your product to inform your company strategy going forward. You can also look at data that is not related to adoption, like one of the biggest things Amplitude looked at last year was which customers were impacted by COVID. Just make sure early, you start to look at the health of your customers and understand how they're using products.
About our guest: Arjun Devgan has been at the forefront of the explosion of SaaS and the evolution of customer success as a function. Arjun began his career in post-sale professional services at NetSuite in 2003 and is currently the Global Head of Customer Success at Amplitude with a team of over 100 professionals across support, success, services, and education.
About our host: Sandhya Hegde is a Partner at Unusual Ventures where she leads seed and Series A investments in SaaS. Before this, she was an EVP at Amplitude leading growth, marketing as well as new products.