Using AI to boost customer acquisition, with IMVU’s VP of Growth, Lomit Patel
How do you grow faster, even in challenging market situations? Oh, and cut the cost of custom acquisition by a factor of three, while also accelerating 100% return on ad spend 5X at the same time?
Well, good data is a baseline requirement. As are insights on what’s working and what’s not.
But AI can help too.
In this episode of Growth Masterminds, we talk to Lomit Patel. Patel is the VP of Growth at IMVU, and previously led performance and digital marketing for Roku. He also ran customer acquisition at Texture, which Apple acquired and turned into Apple News+.
So he knows a few things about growth and customer acquisition. (In fact, as a social app, IMVU is actually growing right now in the coronavirus pandemic.)
He also wrote the book on AI for growth … literally. Patel just published Lean AI: How Innovative Startups Use Artificial Intelligence to Grow in March. It’s full of practical advice on how to use AI to scale growth and user acquisition.
Listen to the podcast:
Key quotes and insights: accelerating customer acquisition with AI
On marketers using AI for growth:
“Less than 5% are really doing this right now.”
Lomit Patel on the three eras of customer acquisition:
- Customer acquisition 1.0: aggregating all your data
- Customer acquisition 2.0: providing clean data to your ad partners
- Customer acquisition 3.0: using AI to automate budget allocation between ad partners
Customer acquisition is the new day trading:
“We’re basically just acting like day traders because we’re seeing bids shifting from any given partner at any given time. And what that enables us to do ultimately is have higher confidence that we can really hit our goals at the end of the day and end of the month.”
On building the AI yourself, or partnering:
“One of the things that we ended up doing was to really kind of do a good and honest audit in terms of where our skill sets really lie at IMVU, and it was pretty clear that we didn’t really have a lot of the core capabilities to really build something like this ourselves.”
What automation enables:
“We … are testing about … a couple of thousand different variations of creatives a month.”
Identifying monetization streams quickly:
IMVU’s AI identifies whether people will monetize via advertising or via in-app purchases within 24-48 hours.
Lomit Patel on what AI has done for IMVU’s customer acquisition:
“We’ve seen our CAC come down over 3X.”
“We’ve seen our ROA, the return on investment, go up over 3X.”
“It used to be close to 5-6 months but now we … get the majority of our recoup of our ad spend within 30 to 35 days.”
And the full transcript: AI-powered customer acquisition
John Koetsier: How do you boost growth with artificial intelligence?
Welcome to Growth Masterminds with John Koetsier. This is the podcast where smart mobile marketers get even smarter. Our guest today is the author of the O’Reilly Media book “Lean AI” which just launched in March. He’s the VP of Growth for IMVU. He’s led performance and digital marketing for Roku. He’s also led customer acquisition at Texture, which Apple acquired and turned into Apple News+.
Lomit Patel, please say hello!
Lomit Patel: Hey, hello everyone. I’m so excited to be here.
John Koetsier: Wonderful. Super happy to have you here. Super exciting, you just launched the book. I have to start here, I mean this is the biggest fact in our lives right now, coronavirus, right? Where are you? Where have you been locked down? How are you keeping up?
IMVU is in coronavirus work-from-home mode, just like everyone else
Lomit Patel: So yeah, coronavirus definitely is the topic of the day every day right now. But for me, we’ve been working from home for the past couple of weeks at IMVU, so we obviously were kind of prepared a little bit to a certain extent, knowing that this was coming down the pipeline. And we’re very fortunate that we’ve been able to work pretty efficiently remotely. So it’s been working out. But as a business, coronavirus hasn’t presented too bad of a problem for IMVU primarily because we’re a social network and it’s another way for people to continue to connect with other users around the world. And for us, it’s definitely been kind of the reverse problem in terms of trying to manage the increased demand as really happened over the last couple of weeks for us.
John Koetsier: Well it’s super interesting that you mentioned that, because I mean as we were just prepping for this show just minutes ago, we were talking about bandwidth issues and my typical audio podcasting platform was not working. You are a social network where people get together with an avatar and have social interactions which we’re desperate and starved for right now. And so you’ve got those issues as well, having enough bandwidth to deliver your service, right?
Lomit Patel: Yes. So for us, you know the challenge is obviously continuing to add more infrastructure into the server capacity because what we’re finding is trying to figure out the right balance between users that are coming in from the US versus users that are coming in globally, and just trying to sort of stagger as much of that demand as possible.
John Koetsier: Wow, wow.
Lomit Patel: And then on top of that, another thing that seems to be very popular with a lot of social networks is live streaming, right? That’s a new feature that’s really taken off and we offer something similar called ‘host rooms’ where people can connect with their friends and host a live event. And so clearly features like that take on increased bandwidth, right?
John Koetsier: Absolutely.
Lomit Patel: Because people are doing that live. So yeah, I think it’s definitely a better problem to have versus the other problem that a lot of people are challenged with right now. But yeah, definitely it’s something that we’ve tried to get more proactive about and something that’s going to continue to get even more, especially as more and more different states, especially in the US and other parts of the country, start implementing more stringent shelter requirements. So it’s going to require more people being isolated, and as we know as humans there’s only so much that you can do alone.
John Koetsier: We are a social species.
Lomit Patel: Yeah we are.
VP of growth: a broad multidisciplinary role
John Koetsier: Absolutely. Maybe let’s kick this off. We’re going to talk about artificial intelligence, we’re going to talk about growth, we’re going to talk about customer acquisition, we’ll talk about user acquisition, all of those things.
But let’s set the stage here for a second, and can you introduce us a little bit to your current role, what you do?
Lomit Patel: Sure, so my name is Lomit Patel and as you mentioned, I’m the vice president of growth.
So primarily my responsibility is managing all of our growth efforts at IMVU. That encompasses everything from acquisition, retention and monetization across the entire user life cycle of … pretty much the easiest way to think about it is how do we bring users in? How do we continue to keep them around? And then how do we figure out how to make some money out of them to pay the bills?
John Koetsier: It’s super interesting that you’re the VP of growth. We released some research recently about CGOs, chief growth officers, and that’s exactly the role that you have, which is this broad role across a lot of what used to be fairly separate, right?
You know somebody brought the customers in, somebody actually built the product, delivered the product that made them hopefully happy, and somebody else worried about how to keep them. You’ve got that full suite within your purview.
Lomit Patel: Yeah. I feel part of the challenge is when you have different groups sort of focusing on different parts of the user journey, what happens is you kind of just run into issues around execution in terms of who’s responsible for which piece of the puzzle.
And one of the benefits of any company that has a head of growth is that your responsibility is to sort of overlook the entire preview of really becoming the biggest advocate for the customer. And helping to evangelize that cross-internally across the whole company to make sure everybody is supporting that function. Because it really is a cross-functional function. It’s not like we have all the resources dedicated to us, so we have to work across all the different groups.
But it’s about ensuring that we’ve always got our eye on the ball—which is about focusing on what are the key projects that are continuing to help move the needle to help the company grow at the end of the day.
Lean AI: AI for growth and customer acquisition
John Koetsier: That makes a ton of sense.
So in all of that, with all the busyness of your job and other things that go along with being a human being, you found time to write a book, and you wrote a book about AI. Talk to me a little bit about why you wrote the book? What did you need to get out there? What did you need to tell people?
Lomit Patel: Sure. So one of the things that I’ve found … we’ve been practicing a lot around AI in automation for the last couple of years at IMVU. It primarily came out of the fact that I really saw this as sort of being a big part of the trend in terms of where growth teams are going to be moving to.
Especially now we have so much velocity of data coming in at us and it’s really hard to really be able to decipher all of that data as quickly as possible to extract insights that you can take actions on, to really be able to differentiate us from our competitors.
And so what I started, what inspired me to write the book was primarily because I’ve seen it have a profound change on our business at IMVU. But I’ve also been speaking about this topic at different conferences and what I found was that there’s not a lot of companies that have really embraced this to the extent that we’ve been doing.
I would say it’s probably a small percentage of the companies, less than 5% are really doing this right now.
And so I felt like there’s definitely a good opportunity to really be able to inspire other people in the growth industry to really be able to leverage AI in automation because ultimately, you know, one thing I like about the mobile growth industry is that we’re so open about sharing and growing together. And so I feel like this is something that could really help us together continue to evangelize and move forward as an industry.
Customer acquisition 1.0, 2.0, and 3.0
John Koetsier: Very, very cool. Now one of the things you talk about is customer acquisition 1.0, you talk about customer acquisition 2.0, and now customer acquisition 3.0. Can you go through the .0’s for us?
Lomit Patel: Sure. So it really sort of goes back to exactly the journey that I had when I started at IMVU. So I started at IMVU over three and a half years ago. And coming in, one of the biggest challenges we had was that we have a lot of user data but the problem was it was living in silos. So we had user data because we’re a cross-platform business. So we had user data from my web business that lived in different servers from the mobile business, that we were trying to get into at the time, that lived in a different place.
And ultimately, it’s really hard to make good decisions when your data lives in silos.
John Koetsier: Yes it is.
Lomit Patel: So a customer acquisition 1.0 is really trying to integrate all your data sources, because unless you integrate the data sources you don’t really have a good preview or a singular view into really understanding your entire customer journey.
And once you have that, the good news is that customer acquisition 2.0 is to really take advantage of a lot of the AI capabilities that already exist with a lot of different partners that most people spend their user acquisition budget with. For example, Google and Facebook, as well as a whole slew of other partners, have continued to make some really good strides and investments into how to leverage AI to enable advertisers like IMVU to get more efficient about helping us hit our goals. But in order to really activate that, you need to provide them with really good clean data signals.
And so customer acquisition 2.0 is once you have all your data integrated then you can start providing the right data to enable them to leverage their AI capabilities to help you hit your goals.
And customer acquisition 3.0 is primarily … you know, the biggest challenge that I saw to us as a business and any other advertiser out there, is how dependent you end up becoming on your partners to tell you how to spend your budget.
And you know, I’ve been in user acquisition for over 20 years and have never ever had a partner ever say to me that we should be spending less.
But on top of that, the challenge is that ultimately I started thinking about is there a better way? And one of the areas where I really sort of got some inspiration from, was really looking at the finance industry, because the way I look at user acquisition teams is ultimately we’re kind of like day traders where we’re always trying to invest the money every day to try and help our companies get a better return on that investment.
And one of the things that the finance industry has done a great job of, and this is when they hired a whole bunch of quants and data scientists, was to build these infrastructures and these intelligent machines to enable them to be able to get better and smarter in terms of how to buy stocks and commodities. And instead of doing it based on humans, it was really based on being purely data-driven.
That kind of inspired me to customer acquisition 3.0. It was to really figure out how we could replicate that because ultimately outside of HR in terms of salaries, user acquisition or growth is really your second biggest spend or line item in the company. So there’s a lot of responsibility that goes into managing the budget and so anything you can do to get better, faster, and smarter around doing that is something that you definitely want to focus on.
And so customer acquisition 3.0 was where we’re taking that same inspiration from the finance industry to really identify now that we have all our data in one place, the biggest advantage that we have over all of these other individual partners that we give our data to is having that singular view on how the business is doing. And that’s the biggest competitive advantage.
So we know exactly how one partner compares to another partner at any given time in any given day, in any given week, in any given month.
And so having that holistic view of the different channels, and the different data, we’re in a much better position to really be able to leverage the key levers. Because the great thing is even though these partners have created a lot of capabilities around AI, they’re still very dependent on certain inputs that come from any user acquisition team which is around bids, budgets, creatives, and goals.
And so these were the things that I started thinking about automating. And the automation was really triggered based on data insights where we had all our data in one place to really tell us at any given time where we should be increasing our budgets, where we should be decreasing our budgets, where we should be shifting that. And the other way to look at it is it’s all about supply and demand at the end of the day. We’ve been very fortunate that everything where we end up spending our budgets now, it’s all programmatically done.
Because that’s what we started finding was that once we were able to identify and look at all of these different partners holistically, we were able to find, at certain times of the day, Facebook could be more efficient than Snapchat, for example, and Google could be more efficient than Apple Search.
But the problem is humanly it’s impossible to really be sitting around 24 hours a day to do this. But if you sort of leveraged the AI capabilities to think like a human and to take predictive actions based on when those opportunities come up, that was the big aha moment for us.
AI: giving marketers control back over customer acquisition
John Koetsier: What I really love about that is that marketers have had this sort of increasing sense of a lack of control as you’ve sort of shoveled money over to Facebook, shoveled money over to Google, shoveled money over to your other partners. And they’ve built their kind of black boxes with lots of AI in there and lots of intelligence in there. And as you said, the more data you feed them the better they are at giving you what you want, absolutely.
But you also lose control and you’re actually feeding the increasing growth or intelligence of that machine inside that black box. What you’ve done is you’ve put a layer of AI on top of that and now you’re saying, hey, you guys get as smart as you want to get. That’s wonderful, that’s great, that’s good for us. We’re going to be smart over all of our sources of customer acquisition and user acquisition and that’s where we’re going to beat our competition.
That’s really interesting.
Lomit Patel: Right. And I think you summarize it clearly because ultimately one of the things that I keep hearing a lot is that we have no control. But the truth is we do have certain elements that we can control and those are the key components that you want to control. Because at the end of the day all of these partners, the biggest thing they want is for you to spend more budget, right?
And you get to control that. It’s all about following the money. You get to control how you want to spend your money. You don’t have to sort of just be completely passive and just hand it off.
So the thing that happens now is every day we’re basically just acting like day traders because we’re seeing bids shifting from any given partner at any given time. And what that enables us to do ultimately is have higher confidence that we can really hit our goals at the end of the day and end of the month.
Controlling the AI that’s managing your growth
John Koetsier: So I have to ask this question. We’ve seen some catastrophes with turning everything over to the machines in the stock markets. Hopefully that’s behind us, but we’ve seen some scenarios where somebody bought, some systems just crashed because everything was going down so they sold, sold, sold, sold, and now they put in artificial brakes to that. And there’s other scenarios we’ve heard of where the machine just made a mistake and billions of dollars evaporated.
How have you put in some safeguards for that yourself?
Lomit Patel: Yeah, so one of the things that we’ve done is very similar to what they do in the finance industry where you kind of have stop-loss orders per se, where you put in certain elements where if something goes up or goes down too much it kind of puts a limit to how much it can do.
And one of the reasons why we ended up doing that, you know, unlike the finance industry, because you can kind of go up and down in wild swings. But the truth is one thing that you want to keep in mind when you do work with all these different AI capabilities of different partners is that you want to be able to make those changes within like a 10-20% swing in any given time.
If you keep making too many wild swings beyond that range you’re going to end up resetting the AI because they’re going to sort of feel the AI is sort of trained to be able to make cumulative changes. But if you do drastic changes then it’s going to sort of feel like it’s a whole new campaign. It’s going to start resetting it and then it’s counterintuitive to the whole program.
John Koetsier: Yes, yes.
Lomit Patel: So that’s what we ended up doing to help.
The technologies IMVU is using for user acquisition
John Koetsier: That makes a ton of sense. Very, very cool. That’s a great overview into what you’re doing. Maybe dive a little bit deeper, what technologies are you using specifically for your AI?
Lomit Patel: So for our AI, primarily there’s three components to it. One is mobile measurement partners and that’s where we get a lot of our mobile data coming in from. A lot of our desktop data comes from our own data warehouse so we use Tableau for extracting a lot of that data. And then all of our post-install user behavior data comes from Leanplum right now so that’s our CRM automation.
And so what we’ve done is pretty much aggregate all three of those sources into one place where we can be able to get that single view, and the way we’re able to do that for us anyway, is in two ways. One is through email address, because everybody who creates an account at IMVU has a unique email address, you can’t create that, you can’t use that email address again.
And another thing is that every user gets a unique customer ID. So those are kind of the two elements that we can use universally to track the user journey regardless of where they come from and then how they end up interacting with us.
Building and training the AI: teams and partners
John Koetsier: Right. Now in terms of teams, did you put together an AI team or did you spread that talent throughout a variety of teams?
Lomit Patel: So what we ended up doing was try and spread it around a variety of teams. But primarily one of the things that we ended up doing was to really kind of do a good and honest audit in terms of where our skill sets really lie at IMVU. And it was pretty clear that we didn’t really have a lot of the core capabilities to really build something like this ourselves.
And so what we ended up doing was to identify a potential partner or SAS platform that we could leverage to help stitch the final piece together for us.
And what we came across was a company called Nectar9. And they were like a small startup that were trying to build something similar. But what they lacked at the time was really getting a lot of user data to really be able to train the algorithms and really be able to perfect the ability of that really working for like a UA context. And so what we ended up doing was … it’s sort of trying to build this because as you know John, trying to build anything is a challenge anyway, but then the other challenge when you try to build something is maintaining it right?
And I didn’t really want to go down that rabbit hole. So I ended up doing this partnership with Nectar9 where we ended up collaborating with them, and I was able to help them through my relationships in the industry, get a lot of these API integrated into their platform from all the major partners like Google, Facebook, Snapchat where we spend and where the majority of people spend their money.
So with those APIs in place, we were able to provide access for them to get all the relevant data signals.
And so they already had the AI capability and the way their capability works is unstructured learning. So it’s primarily getting a bunch of data coming in and then trying to come up with insights based on that. And then based on that, we’re able to start training the algorithm to work for us across looking at things holistically versus looking at each individual partner individually.
And so the best way we did that was to do it step by step.
So we primarily started passing them some data at the time when we were still managing these campaigns semi-manually, with Facebook was one of the first partners we worked with and then we started giving some of the data to Nectar9 and see how the machine would potentially work against the way we were doing things.
And it took probably about 30 to 40 days, but the machine eventually was able to train and get better, and it started being much more efficient than us. And so we started with one small geography, I think it was like the UK or whatever. Then we started expanding it to sort of see if it would continue to hold and it started to hold, so eventually we automated Facebook completely through this. And then we started doing the same thing with Google and Snapchat and a bunch of other partners that we work with.
But it was sequential, it wasn’t just putting everything in there at once.
And in terms of the in-house resources primarily with my team, I kind of changed the skillset because one of the things that I found being a big challenge was really retaining people when it comes to campaign management. Because a lot of people get burnt out doing that work anyway, and there’s always somebody who can offer you $10,000 to $20,000 dollars more to move to another company for that role.
So part of my challenge and frustration was continuing to train a couple of new people coming on, and so what I found was doing it through the machine and doing it once, and then I never had to kind of go for that whole process of building that skillset in the team. And so what I did is I pivoted the skill set that we’d had in a team to really focus on how are we going to support this machine? How are we going to feed the machine, so to speak? Primarily one area where we’ve really made a big stride on and invested in-house is all around creative development.
What one of the people that I have on my team primarily focuses on, on really looking at all the data insights that come in from the machine—to look at what’s working and what’s not working across the different partners that we’re spending money with and continue to work with our in-house team, to continue to keep coming up with new creative iterations that we can continue to keep feeding into the machine.
And we generally are testing now probably about a couple of thousand different variations of creatives a month. And that’s primarily because we’ve realized that’s such a key lever now to really help us tell our story.
John Koetsier: Wow.
AI, retention, and user journeys
Lomit Patel: The other part, the other role that I really upgraded on the team was all around retention and CRM to really leverage all of the data that we’re getting now—but figure out how can we come up with better user journeys that are more personalized to users coming in from different channels to really be able to suit up.
For example, somebody coming in from Google and whatever creative motivated them to come in … we want to be able to provide some continuity on whatever motivated them to come in to keep that going. And then try to get them integrated into really experiencing all the key aha moments on a product as quickly as possible.
Part of that role is really where the AI really helps, is it enables to help us identify what’s the ideal user journey or the behaviors and actions that somebody needs to do, especially around onboarding because it’s key, as you know, in mobile. If you don’t really get people hooked pretty quickly you’re going to lose them.
What we’ve been able to do is identify two different segments because that’s how we monetize users.
One is around in-app purchases and the other one is around advertising. And generally what people do is they end up buying IMVU credits because we have our own currency that people use to redeem against buying all this virtual stuff and creating these virtual worlds. And so there’s going to be certain people that are going to be more likely to spend and buy those credits, and there’s other users who are going to be more likely to take rewarded videos, take online surveys, or interact with advertising to try and earn those credits.
So our AI’s able to identify pretty quickly within like the first 24 to 48 hours based on what people are doing or what they’re not doing to really identify where they are more likely to fall into, whether it’s the in-app purchase path or the advertising path. And based on that, then we start showing them different experiences that talk more to how we know we could end up monetizing those users downstream.
John Koetsier: Interesting. So you’re actually building your user journey, your customer journey in some sense, through your understanding of who they are when they come in.
But you’re also customizing the use of your product based on what you’re understanding, is this user likely to pay something? Are they going to watch an ad or something like that? And so the entire experience is customized to how you’re going to actually monetize that particular cohort I guess.
Lomit Patel: Right. That’s where the part comes in where we need to work cross functionally, like with our product team. And so, these two roles that I have, one of those roles spends a lot of time with a product team focused on how to optimize the product experience to match what we’re looking for. And then the other role spends a lot more time with our marketing team to really focus on all of the creative development and the assets on how we can tell our story better to try and feed all of those back into the machine.
The other role that we have is primarily with our partner because our partners take all of that data and provide the insights. With them we have a chief data scientist that they have, so it helps us because we don’t have to hire this person, but what we do with them is we have weekly meetings at a minimum. But a part of their role is to continue to look at the algorithm and be able to tell us exactly what the algorithm is learning and what kind of outputs it’s kind of coming up with. Because one of the things we want to try to avoid is the biases that that machine’s going to end up doing.
And we want to make sure that whatever algorithms we end up developing are in line with our core value as a business, which is around diversity. So we don’t want to ever become the business that ends up just targeting a certain type of demographic, or gender, or whatever, where over-index is in one area. Because we want to try to keep our user base as diverse as possible.
John Koetsier: That’s super interesting and super smart. Because if the algorithm notices something and there are biases in our society—and there are things that sometimes even unconsciously or subconsciously we do that are resulting in instances of bias, or discrimination, or other things like that—and if your AI learns that, we’re replicating that in the digital environment in ways that are hard to see.
So that’s wonderful that you’re doing that.
I’m also excited to hear that you’re repurposing some of your user acquisition people since the machine does that now, but you’re repurposing those people for higher level jobs which is really, really good.
Key growth metrics: Customer acquisition cost, ROI, and ROAS
You mentioned as well that you’ve got some key metrics that you look at. I wanted to ask you about those key metrics and I wanted to insert a second part of that question and ask you: with what you’ve done with AI, how much better do you feel like you are? And do you have some data around that? Do you have some numbers around that? How much have you improved?
Lomit Patel: Sure. So to first answer the questions around what our key performance indicators are, primarily there’s two goals that we look at on the growth team to really help the business.
One is around cost to acquire a customer or the CAC, and the second goal is around the return on ad spend, but primarily we call it return on investment because one of the big difference being is that when it comes to return on investment we don’t just look at all the revenue debt that we’re able to generate from my ad spend. But we also try to discount, for example, some of the costs that go into supporting that revenue which for us is like 30% rev share that you have to pay Google and Apple on in-app purchases, as well as some of the creator royalty fees.
Because we have a lot of creators, just like YouTube has creators, we have a lot of creators that create a lot of these things.
So we’re kind of like a marketplace where creators are creating a lot of things and then we’re able to match users up with items that potentially would be interesting to them. And whenever somebody ends up buying IMVU credits to purchase these different items, we obviously make a rev share out of that but we have to give a certain percentage back to our wonderful creators. And so our bar is set a little bit higher in terms of discounting those costs.
And then when it comes to cost to acquire a customer, that’s pretty straightforward in terms of how much does it cost for us.
And we define a new customer as somebody who makes a new purchase within the first seven days, so we always look at seven-day cohorts when it comes to that. Just to give you some examples in terms of the results, so just going back to that whole example of the old customer acquisition 1.0 when I first started you know, from where we started with our CAC that used to be fairly high, to where we ended up being at customer acquisition 3.0 with a lot of this AI and automation.
We’ve seen our CAC come down over 3X over that timeframe and we’ve seen our ROI, the return on investment go up over 3X over that timeframe.
Speed of return on investment: a huge advantage
John Koetsier: That’s huge!
Lomit Patel: But the best part about it is ultimately when you talk to a lot of mobile advertisers, one of the other things that people look at is what’s the payback period on your ad spend. When I started, it used to be close to 5 – 6 months but now we … get the majority of our recoup of our ad spend within 30 to 35 days.
John Koetsier: That must have a huge impact on your ability to grow fast.
Lomit Patel: Exactly.
That has really helped us because primarily it helps to minimize the burn rate because whatever we’re spending, we’re able to recoup that money back pretty quickly, so that it ends up becoming like a self virtuous cycle where we’re basically recycling money back and putting it back into growth. That’s really helping us, yeah.
John Koetsier: Wow. Those are great numbers. I mean, those are amazing numbers. If you look at the customer acquisition costs that a lot of startups have, a lot of tech companies have, they’re absolutely massive, including just mobile-only companies they’re absolutely massive. Spending hundreds of thousands of dollars a day on new user acquisition, and if you can take that down by a factor of three … wow … that’s impressive stuff.
I wanted to get in as well, and I know we’re nearing the end of this time that we scheduled together, but I wanted to get in as well some of your thoughts on the future. You’ve done some really, really cool things so far. You’ve got some great metrics.
You’ve had some major improvement, but we’re always looking to the future, right?
The future of AI and growth
We’re always looking at where’s this going to go? Where are you going with AI? And what do you see over the next 6 to 18 months?
Lomit Patel: Yeah, so for me, I feel the future is really going to be about more and more growth teams or user acquisition teams, really pivoting more towards leveraging or building an AI intelligent machine that’s ultimately going to be something that’s going to sit between them and all of these different channels where they end up spending money, whether it’s around acquisition or retention, to really enable them to get better, faster and smarter about using their data, because ultimately it’s all about, you know everybody has data, but the data is of no value unless you can really extract value out of it quickly.
And then the other part to that is the secret to growth is really about just running as many experiments as you can to really figure out what works and what doesn’t work. So you can double down on what does and move away from what doesn’t. And so one of the other things that I feel a lot more companies are going to end up doing is just removing a lot of the human touch points or interactions into the whole process around managing and executing different campaigns.
And it’s going to lean more into these AI intelligent machines to really do a lot of that work for them.
And I feel like companies, for example, Singular and other MMPs out there, they’re basically becoming these warehouses of all this great data for clients. And ultimately I feel that they’re probably going to have some kind of integration put in where there could be these intelligent machines that maybe integrate into these platforms, so that it really becomes a seamless experience to really be able to do it in one place versus what we had to do because we never really had a choice when we started doing this, which was to try and piece all these different pieces together.
Because the problem with trying to put a lot of different pieces together is that there’s always the risk of something breaking in the process. The more you can have it in one place as they say, singular source of data, the easier it is right?
The Lean AI book
John Koetsier: Exactly, exactly. Well, thank you for that and thank you for this time. I want to ask you briefly, if somebody picks up your book, first of all, where can they get it? I assume it’s everywhere, but where can they get it and what is a marketer or a customer acquisition specialist or any executive going to learn from your book?
Lomit Patel: Sure. The book is available at all the major bookstores right now, so from Amazon, Barnes & Noble , Target, Walmart and a whole slew of other places, you can primarily get it on all of those websites. And my goal for writing a book, and it’s kind of written to speak to a couple of different audiences, primarily the whole idea is around it speaks to executives whether you are a founder, or a CMO, or a head of growth, to really help empower you to become a champion for AI in your business and how to do that.
And then it speaks to people that are more kind of at the director level or the manager level when it comes to growth in AI in terms of what are the skill sets that you need to learn to really get better at doing this, and provides a pragmatic roadmap ultimately for any business. Which I know especially right now with everything going on with coronavirus, a lot of companies one of the big questions that they’re going to start asking themselves is, how do we get more? How do we get better at operations? How can we get more done with less? That’s always been a question, but it’s going to be an even bigger question now.
In this book it really provides you a roadmap on how to do that, especially when it comes to one of the biggest spends where companies look to grow, which is around user acquisition. And you know my heart kind of goes out to a lot of companies right now that are kind of going through a lot of challenges with the current coronavirus crisis, but I feel that one of the upsides of any crisis is it forces companies to think differently about any problems that they’re facing.
And so coming back to my story at IMVU, I joined IMVU at a time when growth was really going in the wrong direction and the company really wanted to try something different. And that kind of helped me to really become the champion for bringing this in. And I feel other people that are in a similar role right now could actually use this as an opportunity because it will enable more people to really be open to the idea on how do you leverage AI and automation because ultimately no business is ever going to end up getting through this crisis just by cutting costs, because cutting costs isn’t a strategy for growth.
You have to figure out how you can get better, faster and smarter around using your data to really propel you to start driving acquiring customers. And so I feel the book definitely speaks to that, and you know this time is as good as any to really be able to start having those conversations.
John Koetsier: Well Lomit, that’s great. That’s great advice, I recommend it. I recommend that book and I want to thank you for spending this time with us. I really do appreciate it and thank you for spending this hour or almost an hour with us on Growth Masterminds.
Lomit Patel: Thank you, John. I’ve been a big fan of yours. I love reading your stuff and listen to your podcast, so it’s an honor to be here.
John Koetsier: Thank you so much.
Lomit Patel: Thank you.