The Alchemy of Retention: Forecasting, Velocity and Moments that Matter

Are y'all awake out there? Yeah. At lunch? I know.

We're between you and some drinks. Yep, I'm sorry. We'll get through it together, okay? All right.

So as I mentioned, my name is Dr. Mike Lee. I do have a doctorate, but this is not going to be doctorate level, so it's going to be really fun. I do work at Avalara.

I'm the global head of renewals, but I'm here in a different capacity. I am the founder and chief alchemist of the Alchemy Lab. The Alchemy Lab comes around. We'll talk about this a little bit later, but it kind of came from my idea when I was an individual contributor and even when I became a leader, going into an organization and customer success, sometimes you have to try to figure out where your gaps are in that organization.

How do you close them? I always thought there should have been a list of things that I need to review. That's what the Alchemy Lab does. We create periodic tables for every role in organizations, and we'll talk about it, but we have over 80 periodic tables where you can rate yourself where your gaps are, what's complete, and we create a plan for you to close those gaps over a year.

So I'm really excited about being here. This is one of my favorite topics. Awesome. Well, I am Jasmine Reynolds.

You may have seen me apparently on a slide that I didn't know about until I was in the keynote this morning. I'm doing the Pulse Social, so you found me in the wild, so come talk to me after this. But also, I am the head of customer experience at The Cut, and I am also the CEO and founder of Customer Architecture Group, a consulting group. So we're really excited to be here with you all.

Okay, so I'm going to do a quick poll. Let's grab your phones. This is your moment to be honest with yourself. A hundred and twenty days out.

How confident are you? Oh, we got a couple. A hundred percent confident. Let's talk after this.

Okay. Keep them coming. Okay. Looks like we're kind of swaying towards 50 percent-ish, maybe depending on the account, maybe depending on the relationship, which is fair.

And then there's the other half of us that are like, "Yeah, I'm throwing spaghetti at the wall. I have no idea what I'm doing," which would make sense why you decided to come to this session. And thank you for being honest. This is a safe space for all of us, so we appreciate that honesty.

So really quickly, show of hands, are there any ICs in the room? Okay. Okay. Okay.

See you in the back. Perfect. All right. So basically, I want to talk to the ICs just for a moment, right?

Because you're the ones carrying this, the forecasting piece of it, and you've been asked the forecast renewals, but if you're really honest, it just feels like a guessing game, and I've been there, right? You're doing the math on vibes, so you're forecasting on vibes, and you're talking to the customer in that QDR, and reading their tone, and relaying that call in your head. You're thinking to yourself, "Oh, I think this was good. I think that was a good conversation," and then that turns into some kind of color, red or green or somewhere in the middle, right?

So you're just kind of fumbling through it, and we get it. So here's where we want to be really clear, okay? Customer success was never built to forecast revenue. In most cases, a lot of companies are afraid to give customer success teams revenue ownership, but we are asked to forecast it.

Interesting concept, right? It was built on managing customer relationships. So we do things like health scores, activity logs, NPS scores, QBRs, and all of that is helpful to an extent for relationship management in itself, and none of it was actually designed to answer a CFO's question of, "What's our forecast? What's going to close?

What are we confident about closing?" So let me make this a little more real for you. Think about the last time you realized you were hungry, and we're getting into the leading versus lagging, right? So you didn't just wake up out of nowhere and say, "I'm starving." There were signals. You skipped breakfast, maybe even lunch, because you had back-to-back meetings.

Your energy dipped. Those are all leading indicators, and when you finally said, "I'm starving," that is lagging. That's the outcome. CS operates almost entirely in the "I'm starving," and forecasting requires you to live in the signals before the moment.

Let me repeat that. Forecasting requires you to live in the signals before the moment. That's what real forecasting is. So why does this keep breaking?

Because every forecasting number passes through three gaps, and by the time leadership sees it, it looks like it's a fact. They believe you. You've sold them a dream. Gap number one is communication.

Your customer tells you that whatever they're comfortable telling you, they're your friend. Y'all are buddies, right? Y'all have a good relationship. Okay?

When you're walking into a QBR with them, they're not going to immediately tell you, "Hey, our budget just got cut this morning, just found out," or, "I'm looking for a job, and I don't even plan on being here next month, so good luck, but I'll tell you whatever you need to know." They perform health, not because they're lying to you, but because they are managing you just as much as you are managing them in that relationship. Gap number two is interpretation. You take what they said, and you filter it in your head because of, again, that relationship. So through optimism, through the relationship loyalty, through whatever you need to tell yourself for that account to be okay, because sometimes we're not always honest with ourselves.

And now it's already drifted from the truth. And then gap three is translation and interpretation into a field or a score or a color. So your CFO is like, "Hey, what's our forecast?" And you're like, "Oh, it's green." That tells him nothing. It's not helpful.

So in my experience, I have always looked at these forecasting spreadsheets, and when you look at those columns, you're seeing things like sentiment, customer health score, when was the last QBR? Hopefully some of these things you also see on your forecasting spreadsheet, but that doesn't, again, help you forecast. You're looking in the rear view mirror when you should be looking through the windshield. So let's call it what it is.

This is not a forecasting problem. It's a human architecture problem. And you can't fix a human architecture problem with a better spreadsheet. So I think, and one thing that I've personally lived by is you attack the problem, not the person.

The better equip are CSMs, or ICs in the room, on how to do this properly. Sales has learned this walking in the door. They know how to forecast. Now, sometimes it can be a little shaky, but CSMs are overachievers.

We can figure this out quickly. A real forecast has a very different standard. A CFO doesn't consider something a forecast unless it answers four things. How much revenue is at risk?

When will it happen? What is the probability? And what can we do to change it? Sales, again, does this every single day.

CS almost never does. And if you figured that out, we'd love to hear more about your experience with that. Okay. So what does this look like in practice?

You need three layers. Layer one is revenue. The numbers, the renewal, the expansion, the churn, that's where most teams stop. Layer two is momentum.

This is where it honestly gets real. You start to look at adoption velocity, executive engagement, workflow integration. Is this product actually embedded, or is it just present? Layer three is intervention.

What are we actually doing about it? Now what do we do to make that actually happen, and what are we doing about it, and the actions that are in motion to change the outcome? Forecasting is not just a number. It's revenue, momentum, and intervention.

And if one of those are missing, you've already missed the ball. I'll give a personal experience. I had one of our largest accounts at a previous employer. And this account, I mean, I met with them literally every single week, sometimes multiple times a week.

Fantastic relationship with the decision maker, with other decision makers, having constant conversations. And it wasn't until the renewal happened where it came out of what felt like left field that they told me, "Oh, we might not renew." All of that work and all of those things, I'm like, "I did a QBR every quarter. I taught my consistency." What talking with them was spot on. Everything felt good.

But there were things that were hidden that I was unaware of, and I realized that we were just forecasting wrong. So now that we've got clarity on the problem, we've named what's broken, and we definitely can understand what good should look like. There's one signal, one number that separates guessing from knowing. And in this world, in our world, it's velocity.

Mike, why don't you take it from here? All right. So first, let's talk about velocity, what velocity actually is. Because in the SaaS world, it can mean a couple of things.

If anyone here that has experience with agile, velocity is the amount of points a development team could complete in a sprint. We're not really talking about that. Velocity is motion. And there's a concept of relative velocity.

So if you're flying in a plane, say two jet airplanes, I'm a big airplane fan, two F-15s, F-16s, something, they're flying next to each other, the relative velocity between them is zero. But that doesn't mean they're not moving, right? But if one happens to turn away, the relative velocity going away increases. If it's coming together, the relative velocity decreases here.

So it's about movement. So let's keep that in mind as we start to introduce adoption velocity, what that actually means. Now I want you to take a moment to look at this. Don't worry about taking a picture of it.

You're going to see a QR code at the end. You all have access to this on the alchemy lab on our website. There's actually QR codes here as well, kind of scattered around. But at the alchemy lab, what we've done is, like I mentioned a little bit earlier, we've built over 80 periodic tables for each function in a SAS organization.

And even more than just SAS, like I'm an elected official, I created one for a county manager, a city manager, because every role has these elements. And so each element represents a part of a role that can't be broken down any further, just like in chemistry. Everyone here should have at least seen the periodic table, even if you're from different countries. Now this is not something for you to memorize.

I remember in high school when they showed the periodic table and we had to memorize it. I lost it. But this is not for you to memorize. But this is what the alchemy lab believes is what retention means.

This is not a specific role. This is more of a function or a capability. So that's what we believe it is when you put it all on the table. 65 elements, 65 distinct things your organization must have a point of view on to run a sound retention motion.

Most of us are managing maybe 10, maybe. But just so you know, we got a group of CCOs and VPs, a couple CSMs in there to review these elements. And we actually just changed this to be 102 elements. If you start to add in AI and some data things and so forth.

So when you scan the QR code and you log into the alchemy lab, you're going to see the full periodic table and you'll be able to see what each element means and what good looks like on each one. You can rate yourself, create a project plan to close your gaps. Now I'm not showing this to you to overwhelm you. I'm showing you this to give you permission.

Permission to stop pretending that adoption is a health score or any other, some of the time to value those kinds of things. They're not. But I want to focus on just one element today. One that most of us maybe never have named out loud.

And that is velocity. This is the 59th element on this chart anyway. It's in the group data. They have all different groups you can see.

And so we're going to focus on what this actually means as it relates to how we can improve our forecasting. But before we go further, I need to draw a very, very clear line here because these two things are not the same. Time to value is a point in time. It's a moment.

60 days, 60 days, 90 days. When a customer sees some outcome of using your product for the first time, they get a report. They said they closed a loop. They saw some kind of result.

And they said, oh, this works. We all know we want that to happen as soon as fast as possible. And I'm no different when I was a CSM. Our first CSM role was at a place called Aptio.

And we fought for that first value. It was budgeting. It has to do with budgeting. And when they ran their first budget, it was great.

They love it. We're good to go. But that's what we're kind of taught. We're taught that it's table stakes.

But here's what nobody's asking. What happens next after time to value? Nobody is not a point in time. It's a direction.

It's the answer to after the first moment of value or what I think we call the moment of truth back before I say BC, but before COVID, I was an individual contributor before COVID. So it felt that long ago. What happens after that? After the first moment of value.

Are they speeding up? Are they leveling off? Are they slowing down? Those three words, accelerating, plateauing, and declining are more valuable to your forecast model than anything in your CRM.

So what do I mean by that? Here's the uncomfortable truth. Most CS teams measure time to value. You hit a milestone, check a box, move on.

But somewhere in the back of your mind, you tell yourself, they got value, they're going to stick around. But you assumed. You assumed that because it worked once, they fell in love with it and that would carry them forward on its own. But that assumption is where churn lives.

That's where it hides. Your customer continued using the product. They remember what it felt like. They stopped because something changed.

In their team, in their priorities, and in their usage pattern. And nobody saw the signal because nobody was watching what they did. If you think about what Jasmine was talking about her, that account where she was talking to them. They were loving it and everything.

But maybe she was not seeing or maybe the company didn't have signals to look at and what they were actually doing. But that's what velocity tells you. It's the color you were missing after your time of value moment. So let's talk about accelerating plateau and declining.

What that actually looks like in real life. Now obviously if we're accelerating, that's a good signal. But what does accelerating mean? That means not just the surface level logging in.

That means maybe their licenses are starting to, they're assigning their licenses. Everyone that's getting their licenses logging in. Maybe not just running or using one feature. Because think about this.

If everybody uses, if they're using all their licenses, they're increasing the assignment of their licenses. That's accelerating. But what happens if they're only using one feature of your product? That's it.

So it may look good. But they're only using one piece. Accelerating should be the expansion of the features, expansion of the usage within your product. You had to run it out of report.

Well, let me give you an example. At Conga, I worked at Conga for a while. Anybody here heard of Conga? It's actually a pretty cool company.

But it was a company called Aptis before and it was really just about contracts. And what we would look at is not just if the lawyers were in there creating the contracts, but what the different types of contracts. Were they using our sign-in product or were they going out the docu-sign? Are their contracts being viewed by their customers?

Those types of things. Not just are they creating contracts and that's it. We wanted them to expand throughout their uses. We wanted them to use the different parts of the application.

And if they're not doing that, maybe they're not accelerating. Plateauing. Now, plateauing may not be a bad thing. It may be a signal for a CSM to check in.

So say a customer has used all their licenses. That's all they really need. And everybody's just kind of moving along or kind of doing it. They're using multiple products and so forth.

It may just be that they're actually at the place where they're fully taken into your product. That could be a signal for growth. Hey, you know, you're going to have all your licenses assigned. Let's talk about expanding what you're doing.

Let's talk about some other features. Let's talk about some upsells, cross-sells, so forth. And then obviously declining. That is not really a positive spin you could put on that.

That means they were going upward, but now they're kind of trending down. Yeah, they saw time to value. Like we saw on the previous slide, you know, the point in time, time to value. But since that point, maybe someone, maybe a group of them in their licenses stop logging in.

They're not using the product. They're not using the reports that were customly done for them. That's what declining looks like. And that's obviously a signal.

As you say, hey, we may want to not count on the new rule. Let me get engaged. I'll get engaged and we'll send something to the CSMs to get them engaged really quickly. But hey, we may not want to.

So the key here is the different levels here sends different things to your CFO. And that's what your CFO wants to know. They want to know what are the chances. And I believe, I personally believe, accelerating is probably a better chance.

Platoon, it's okay. We can, like in sales, they say, you know, it's the 50% shot or a 70% shot. Declining is at 10%. And that's what that's, we have to learn how to talk to our CFOs.

And that's how they want it. That's what they want to talk about. So velocity is the color we're missing. That's what mentioned before.

We assumed that they fell in love with it. Everybody has the best product. I mean, I think the Alchemy Lab is probably the best thing that hit SAS ever. I mean, we haven't launched yet, but still.

But that may not be the case. We need something to tell us what they're actually feeling. Everyone wants to be your best friend. It's hard to break up.

So like, when I was in high school, we didn't have cell phones or stuff. You couldn't text a breakup. You had to actually face-to-face break up with somebody. That's hard to do.

So those who may be younger, you guys got it easy. But I'm just saying, we have to understand what is actually happening in the account. So it's really quickly. I want to talk about what we're missing here.

So one, go look at how you're defining adoption. Not just licenses activated, but feature engagement, engagement frequency, depth of use, time and spending the product. That's the raw material. If you can't see that data today, that's your first project when you get back on Monday.

On a weekly or bi-weekly basis, tag each account for the direction, accelerating, plateauing, and declining. The three buckets, the accounts that are accelerating, carry high confidence in your forecast, accounts that are plateauing, especially with unused licenses, get a flag. Accounts that are declining will have a CSM assigned to them this week, or if they have a CSM, they're going to engage this week. And then take the three, take that directional signal into your forecasting conversation, not as color, as data.

Your CFO doesn't want to hear what we think. Well, doesn't want to hear, we think the customer's healthy. They want to hear this cohort is accelerating, so our Q3 expansion confidence is 87%. This is a completely different decision.

So your customers are moving. They're going somewhere, and all of them are. The question is whether you're watching. Velocity is how you watch, and that's what we're here presenting you today, is a new signal, velocity, how to look at things a little bit differently.

And when you do, your CFO will be very happy with you. Now everyone here, I'm sure, has had forecasting experiences, good and bad, and we're here in these really nice lab coats. So we're going to have us a little therapy. We're going to have a little therapy, and we want to understand how have you all fared with forecasting?

We have some questions here and all. Yes. Come back up. Now we're good.

Okay. We do have some questions. First one, where do leaders most often misread adoption velocity, and what context do they need before acting on it? I'll start with this one here.

I kind of touched on it a little bit when I was talking about acceleration. I think people are very quick to say when they see a signal, say, "Oh yeah, that's it. They're accelerating. Good to go.

CFO, count on it." But you're not looking at the full picture, because they may be, like I said, they may be accelerating in license usage, but what if everyone's only using one piece of the product? They're only exporting your product for a PDF when they can actually manipulate things, and they can grow things, and change it, and so forth. So you want to look at the full picture, because seeing something accelerate may not be the full picture. So before you mark that as a signal, make sure you have a more complete picture.

Yeah, I'll add a little more color, but exactly what you said. I think thinking outside of the picture that you've already painted, because right now, what you're doing is you're looking at, all right, so time to value, there's value there. You assume that that's going to take them through. You need to extend your, like the path, the journey beyond that moment.

That actually may look slightly different. There may be different paths, so you need to look at the multiple paths that may happen after value, because it may be, like he said, they're using on their seats, not using all the features, but another path could be using it in a way that they hadn't even thought about yet, right? So you walk in the door saying, "Oh, this is what I want to get out of your tool," but you need to be innovative and consultative and help them understand other ways and maybe even other departments that could use your tool. And then there's another, again, that's another path.

Yeah. Thanks. Where do leaders, how do you apply velocity when you have very little usage data? Well, so I think velocity, I think anything that happens after the first moment of truth, first time to value, is something, right?

Now very little usage data. I'm assuming that usage data means logging in or different parts of the product and so forth, but the lack of usage, usage data is a signal in itself, right? What does that mean? So usage, this is going to get a little technical, but a while ago I was trying to figure out what usage actually means to a customer, to a company based on a customer.

And so you may have a customer, you may have say a thousand customers and you see a number of logins, right? Say a hundred. Just say, "I don't know if that's good or bad, a hundred." Well, we don't really know what that means for that customer, right? And so I came up with a super complicated calculation.

It worked, but it was really averaging what logging in or usage meant for individual customers. So if you get this average and in a particular month, there were two standard deviations above what they usually do, their usage is up, right? If it's two standard deviations below, that means their usage is low and that should give you a signal itself. But if it's right in the middle, it's just kind of average.

So if you have little usage data, there should be other signals that you can look for that may align with what velocity means, right? So they may log in once a week, but what are they doing when they log in? How long are they staying on your site? Are they using the features that you need them to use or that they need to use or why they purchased it?

So little usage data may not be a bad thing, but it depends on what else is happening. Are they actually using your product in the different ways that it's designed or that you represent them? That's what I think, Mary. What do you think, Jerm?

I agree. I would say that outcomes maybe even would be more valuable, right? So if you don't have the usage data, how close are they to their outcomes or how far away are they from it and track that? Have you ever seen velocity mask client frustration?

For example, maybe a time-consuming work around has caused the client's usage to increase. That's a good question. I think that again comes with, well, do you want to take this one first? Nope.

I don't want it. I don't want it. You got it. Well, so I think in this situation, again, an increase in logins or so forth is one thing, but what are they doing?

Why are they logging on? Are they just logging in to check a box each time they say, "I logged in," so their leadership says, "Oh, yeah, logging in," or are they in there actually doing the work? So I don't think, I mean, I think the time-consuming piece or the usage or logins, I think it balances itself out if you start to look at what they're doing when they're logging in, the other signals. Because again, usage is usage.

Some people are going to log in because they love your product or it actually makes a big difference in their lives. Some people log in, they'll do the work and log back out. That's just kind of, it kind of washes out. It kind of averages out.

But the key here in talking about velocity is not necessarily just looking at one thing as looking at the signal in general. What all are they doing? What other pieces of the product are they using? Are they increasing users?

Are you seeing them? Are you seeing the company going through their enablement or their training to use different parts of it to enhance what they're providing themselves or organization? That's what that looks like and it may be time-consuming inside, but what are they doing while they're there? That's what I think we have to take a look at with velocity.

If some features have cyclical usage, how would you think about accounting or normalizing this change? Simple meaning seasonal kind of thing. Is the person in here? Can you expand on it?

There's three up votes there. But I'm going to... Maybe less. They were anonymous.

We're talking about yearly. I'm going to give you a general example. It was either a customer and our usage spikes at movement and our small sector. So obviously September is super high and then it'll drop back down and that's not necessarily a negative percentage.

That's just the seasonal power class. She said she works for a company that's, I guess, with universities and her usage spikes and move in time where kids move in August, September and it kind of balances out. But that's your pattern. That's what I was going to say too.

That's your pattern. So that dip in usage doesn't necessarily mean that they're not getting value out of it or they're not using different products or the students that are logging in and so forth. That's not negative. That's just what your usage is.

Yeah. I was going to say that your velocity is at the top of that curve and then maybe even how fast you get up to that curve or they prepare to get to that curve. So that's where I would kind of measure that velocity. It's an opportunity to start that a little earlier.

I don't know, but yeah. That's where my thought went. That was my question. Okay.

You got some thought votes. Yeah. I know. That was good.

Definitely. All right. How do you get executives to trust leading indicators when they're used to relying on lagging metrics like renewal rate or GRR? Yeah.

That's why we're having therapy sessions like the trigger. That's right. I think it's articulating the value and paying attention what happens before that. The value is actually in what before you start to look at that GRR number or that renewal rate.

I get that they're more outcome focused. That's okay. We'll share those numbers, but I think velocity is a nice addition to that rather than the account is healthy. Yeah.

Do you have any tips for leveraging AI to track velocity? Yeah. I mean, actually, I was sitting in the keynote this morning and they were talking about like an agent and they had adoption, like this adoption agent. Adoption is really the increase of usage within your product.

I believe I was sitting there, I took a picture of it and I sent it to my VP. I was like, "I think we can use this." Not my VP of the alchemy lab, because there's none of that. It's VP of Avalara. I was like, "I think we can do that to understand what they're adopting and how deeply they're using it." Right?

Yeah. I think you're going to see a lot of agents to be able to do that. If you start to identify different pieces of your product that you really want to signal on if people are using it or how they're using it, that data is going to come out raw. You can use AI to really understand how deep and how far they're using your product for adoption itself.

I absolutely believe we would be able to do that, but first you're going to have to tag. You're going to have to figure out on a whiteboard what usage or adoption actually means for our product with these different pieces. Once you're able to tag those pieces, you'll be able to use AI to understand what the trends are. I believe.

Yep. How do you handle adoption when there isn't one product they subscribe to? In example, our largest client could have up to 15 different products. Do you focus more so on overall subscription utilization then?

I would definitely say the overall subscription utilization is what I would focus on in a situation like that. Maybe even, I don't know if the different products, if the adoption would be consistent if there's 15 products, but maybe if there's more heavy adoption in one or five of the 15, it's something that I would pay a little more closely to. That's where my thought went. Avalara has a ton of products.

Somebody said, "Who has?" Avalara has a lot of products. We do look at overall usage, but we use signals to say, "All right, this customer doesn't really use Avalara Tax Research. They don't really use it much." When renewal comes around, we say, "Well, how about we switch this product out for a different product?" Just because they're not using one product, but they have 10 products, just because they're not using one of them doesn't mean they're not healthy. That just means that maybe it's just been one that a salesperson threw in there and said, "Hey, put that on there.

You all know how sales works." Did that answer your question? Yeah. She was just saying that sometimes customers need certain products at certain times in their subscription and sometimes they don't need it. That's what she was saying there for those who couldn't hear.

I think we have time for one more. How do you get CSMs to forecast velocity with clients when there is a large account to CSM ratio, especially if automated metrics cannot be built in 90 days? AI, sorry. This is a RevOps track, right?

Get the work. Hey, okay. This is RevOps and CSOps, right? No, I mean, so I get that, but it doesn't have to be perfect at first.

It can be very general. It can be one thing. How are they using this? You start to add more and more and more.

It should be a canvas. You should be able to paint slowly at a time. My daughter wants to be an artist, and I tell you, it's not going to be perfect when you first put your pencil on the paper. Just start.

Start with one. Start measuring what they do in a particular part of your product and then start to add the different pieces. Look at what your customer is doing and then start to build out. Of course, leverage your RevOps and CSOps people.

This is what they're here for. This is what they want to do. These are real world. Don't be afraid to iterate.

Just put something out there and keep tweaking. All right. Just a couple notes quick there. You can leave feedback for the session.

If you go into the agenda, click on the session. There should be a survey option. We appreciate that. Don't forget about the Pulse Party tonight at the Vanderpump Hotel.

Maybe check on that in case I have it wrong. Okay. Perfect. That's at 7 p.m.

tonight. Looking forward to seeing everyone there. Please give a round of applause for Jasmine and Dr. Mike.

Can we switch to the last slide there? If you're interested in retention alchemy. Yeah, there's one more slide. You can- QR codes.

There's these little QR codes around here. Someone asked about the formula. I wrote a white paper on adoption velocity and there's a little QR code to get to that here. It's on the chairs, some of the chairs.

If you want to read more, I'm a doctor. I have a doctor. I'm going to write about this kind of stuff. Makes me happy.

You can log in and you'll have the access to retention alchemy. You can see it all and use it fully. Thank you all.