If I joined your company tomorrow to lead growth, I’d run a specific 6-step playbook. Without it, I’d basically be stuck guessing where the leverage is or why the numbers haven’t moved. That’s time I should be spending in the most valuable areas of your growth model instead.
So I’m going to walk you through the 6 steps - and the tools that helped at each one.
Each step has a tool behind it, and the tools feed each other. At the end, I’ll also share 2 changes I’d make if I rebuilt this stack today.
If you’re new here, my name is Andrew. I’ve worked in and around tech startups for the past 15 years. I’ve been head of growth twice for 2 different high-volume scaleups, and I’ve coached 100+ directors, VPs, and heads of growth since going out on my own in 2021.
These tools fit into my growth process but they’re certainly not the only tools.
None of these companies paid for my endorsement (though I do advise one of them, which I’ll disclose below).
Prefer to watch? Here’s the full video. Or keep scrolling to dig in.
1. Find the leverage
If I was hired to lead growth at your company, my very first step would be figuring out where the leverage is. Job number 1 as a head of growth is figuring out how we grow faster.
To start doing that, I need some type of analytics tool. Historically, I’ve used Mixpanel.
If you can’t see where users are dropping off, it’s impossible to prioritize. Every idea in the backlog or on the roadmap looks equally good.
Then later, after you’ve been executing a bunch of this work, your CEO or one of your executives asks, “Hey, what impact did [that thing you just shipped] have on the business?” And you can’t make the connection, you’re stuck talking about localized data like click-through rates or engagement instead of business impact.
We used Mixpanel at Wistia. It helps you build your own funnel reports and drill-down views based on what users do in your product, and it integrates with your A/B testing tools (I’ll get to my favorites in a second). That meant my team and I spent hours inside those funnel reports. We tracked all of our growth model data in there, all of our product engagement data, and all of our new user onboarding and activation data.
We could connect how the things we did on the website trickled into product behavior and eventually conversion down the road. That was super helpful, because we used it for a lot of our pricing tests.
In a product-led company, data is an incredibly important currency.
When you can connect a change you made to the impact on the actual business, whether that’s number of customers or revenue, leadership likes that. Your execs understand what you’re doing and why. And your prioritization and alignment, which is a big part of the success of this work, gets so much easier.
Our business intelligence team owned the Mixpanel account, but our growth team basically lived in it.
If you’re at a new startup today with the luxury of picking a tool off the shelf, there are a couple of other options to consider.
A bunch of folks in my coaching portfolio are building their own dashboards and doing all their analysis in Claude. They use something like Segment for event tracking and pipe the data wherever they need it for visualization and analysis. I think that’s a good starting point. I’m also intrigued by PostHog, which I’ll talk more about later.
But job number 1 as a head of growth is understanding how to grow the business faster, and we can’t do that without understanding where the leverage is.
2. Figure out why
Once we’ve found where the leverage is, let’s say there’s an area of our growth model that’s underperforming compared to benchmarks or what we think it could be doing.
The second step is to dig into why.
A dashboarding tool like Mixpanel shows you where your conversion rates might be low, but it doesn’t tell you why.
It doesn’t tell you see where the friction is. And when you’re working on something like new user onboarding or activation, you can’t guess. Guessing is time-intensive, labor-intensive, and expensive when you think about the cost of not fixing it.
I liked using FullStory session recordings, so we could watch exactly what users do or don’t do inside the product.
We built a whole ritual around it with my team, called FullStory Fridays.
We’d get the team in a room and watch 10 or 20 real user sessions of people navigating whatever area of the growth model we were focused on.
We’d walk out of those sessions with a laundry list of impactful changes: bugs we didn’t realize we needed to fix, UX changes, empty state designs, flows, and some very big tests we might want to bet on down the road.
During one of those sessions, we saw something pretty impactful.
We knew from the Mixpanel data that about half of the people who signed up for Wistia never uploaded a video. Without a video, you can’t really get much value from the product, and you’re probably not going to convert into a paying customer.
So we had a sample video people could borrow from us. It starred Lenny, the office dog, walking around being cute with an old VHS tape in his mouth. Sort of an nod to Wistia being a new way to do video.
What we didn’t see in the Mixpanel data was that basically every person who clicked play on that borrowed video watched the whole thing.
When we saw that in FullStory, we asked ourselves a question….
What if the video we showed them, instead of just being cute and brand-y, was educational? What if it showed them what to do inside the tool while they watched it?
So we tested it.
We made another version of the video that said, “Hey, this video lives on your media page. Go up here and look at that. Then go over here.” And in Fullstory, we watched people literally follow along on their screen as the video told them what to do.
It was pretty cool. Activation in that test cohort went up by about 30%, from something we never would have seen in Mixpanel.
There’s a lot of functionality in that tool, but growth lived in the session replays. It’s still a big part of my personal tech stack today.
If you just need replays, other tools can do this.
VWO, which I’ll talk about in a second, has it built in. So does PostHog. Hotjar has something similar. But if you’re trying to find the actual friction points (and what might help you break through them), having some way to watch users navigate the tool on their own is incredibly helpful.
3. De-risk the fix
Even once I understand the problem in more detail, before I ship something, especially a high-stakes test, I want to de-risk it. I want to double-click on the problem and see if I can pre-validate that my solution is likely to work.
That’s where user testing comes in.
Watching shows you what users do. Asking helps you understand why they take an action or don’t. There’s no better way to understand what’s stopping users from taking an action than asking them.
My teams used UserTesting.com for years.
You give them your website or your app, show it to strangers who match parameters you set, and give them a specific task: go through this set of actions and say out loud what you think you’re supposed to do.
We did this a ton for pricing and packaging, and also for our new user onboarding tests. We’d ask lots of open-ended questions.
On pricing and packaging:
Would you purchase one of these plans? Why or why not?
Which of these plans do you think would be best for you?
If you didn’t move forward with one of these plans, what would be the most likely reason why?
On new user onboarding:
What do you think you’re supposed to do here?
What do you think the next step is?
What would stop you from taking the next step? Or what almost stopped you?
What do you think the main value of this tool is?
All of those are incredibly eye-opening questions that help you pre-validate the work you’re about to ship.
One of the coolest uses of user testing I had was during the pricing and packaging work.
We’d done a lot of modeling to figure out the most effective new pricing and packaging for Wistia. We were pretty confident the team landed on the right solution, but we didn’t know the right way to present it to users.
So our designer created a few versions of the same pricing page.
Same price, same feature lists in each plan. But different ways of presenting the options.
Version 1 said 10 videos were included for that $100. When we asked folks if they thought the plan was valuable, or which one would be the best fit, they all thought the same thing out loud: “Wow, 10 videos for 100 bucks a month. So that’s $10 a video? I’ve got a couple videos that are just internal, and one isn’t even edited. I don’t know if we want to pay 10 bucks a month indefinitely for that stuff.”
Version 2 was the exact same plan at the same price, but it said your first 10 videos were included for free. Every single person we showed that version to said, “Oh wow, that’s a pretty good feature bundle for just 100 bucks.”
Same price, same features, totally different perception.
An A/B test probably would’ve shown us that version 2 won statistically, but it would have taken weeks to collect that data and we never would’ve never understood why. That means we couldn’t have taken the learning and applied it to other areas of our growth model.
That’s what user testing gets you. You de-risk the decision before it get shipped. I still recommend this to all my coaching clients.
Unfortunately, UserTesting’s prices have gone up dramatically. Today I’d probably recommend Userbrain. That’s what we used at Postscript, and it’s a little more startup/ scaleup friendly on price.
It’s so helpful to have people who match your ICP review your site, your product, and your user experience, answer open-ended questions, and then use what you learn to guide your thinking before you run a test.
4. Test the fix
Once I’ve pre-validated my solution with user testing, at some point I want to implement the version I’m moving forward with and see how it performs against our baseline.
For that, I’d run a test using a tool like VWO (Visual Website Optimizer).
A lot of startups and scaleups start with free a/b testing tools. But once you’re in the optimization stage of growth, you’re ready to test, personalize, and break through your growth plateaus. You need rigor. You need statistical validity. You need a tool the team can use to ship quickly, with all your experimentation tracking built in.
I actually grew up using Optimizely, but at some point Optimizely went upmarket and priced out startups and scaleups. So at Postscript, I used VWO.
There are a few tests we ran with VWO at Postscript that I’m really proud of.
Postscript is an app in the Shopify ecosystem and you could install it from 2 places, either directly from the website or from the Shopify App Store.
My instinct as growth lead was to keep people where they were.
IE, if visitors were on the website, they should install from the website. If they were in the App Store, install from the App Store.
But in our FullStory data, we saw users on the website hovering over the App Store language (on our homepage) and trying to click it. So we tested letting users go from our website to the App Store, and linking from the App Store back to the website.
I thought signups would probably go down and we’d lose users in the process. Conventional wisdom says that on the signup page you remove all the navigation and let people do only 1 thing: move forward, or eject and go backwards.
Instead, total signups went up by about 20%.
That was a complete surprise, and it changed how we thought about those 2 channels. They were working together.
Testing is a huge lever to break through your growth plateaus, and a tool helps enables those tests to be done accurately.
Optimizely is no longer a fit for the startup and scaleup community, and VWO is a great fit. One of my other clients has been on GrowthBook for the past year and seems really happy with it, and the price is solid. PostHog lets you run experiments too.
Any of those would be a great fit for anybody leading growth today.
5. Nudge users inside the product
The next step in my growth process would be nudging users inside the product toward the next action - helping them get stickier, increasing retention, and combating the churn challenge with upsells and cross-sells.
In a product-led company, your product is a huge channel for communication, conversion, and engagement. You need a way to maximize that channel.
At Wistia, we custom built our own checklists, popups, and communication tools. It worked amazingly, but it took forever. And over time I learned about something called code decay, which I’d never heard of before.
Basically, you ship a piece of code, and at some point it goes bad and bugs pop up because of other code changes around it. We didn’t have anyone supporting those things full-time, and eventually that happened to us.
So at Postscript, I pushed for us to buy a third-party low-code or no-code tool. We bought Pendo and were pretty happy with it.
It let us ship guides, checklists, and even customized onboarding recipes that routed people based on what we knew about them, nudging them toward activation, retention, and habit building.
In a product-led environment you need ways to maximize the product communication and conversion channel, and you need a tool to do that.
6. Follow up outside the product
The last step is following up with people outside the product. Because in-product nudges only reach people who log in.
For everyone else, you need a way to take what they did (or didn’t do) in the product and use it to trigger the right message, pulling them back toward the next action.
To do that, I need a tool that connects product data to your CRM and automates the follow-up. For me, that’s HubSpot.
It’s worth flagging that I worked at HubSpot for 4 years. So please know I’m biased here.
You can send all your product data, milestones, and CRM data into HubSpot, then build campaigns based on what people did or didn’t do. That’s what I did at both companies.
The most impactful campaigns were segmented programmatic campaigns that nudged existing users to upgrade from free to paid, or from Pro to Enterprise.
We identified product-qualified lead signals and used them as triggers to kick off evergreen campaigns that showed people features and encouraged them to take the next step to the next plan.
Those were some of the coolest campaigns I got to work on. We obviously had onboarding flows and plenty of non-product emails going out too.
There are some newer tools in this market (and some folks claiming to have built their own tools using Claude for this). But you need something to communicate with customers outside the product, using your product triggers to nudge them toward the next action.
2 changes I’d make if I rebuilt this stack today
That’s the 6-step system I’d run if you hired me to lead growth at your company, and the tools I’ve used at each step.
It worked. But if I was rebuilding it from scratch at a startup with no history and no existing tech stack in place, I’d keep the steps and make a couple of changes to the tools.
The first replaces part of the growth playbook we built ourselves at both companies. The second lets you cover half of the stack I just talked about with 1 tool.
Let the product sell itself
The first change is that I’d use the product to market and sell itself more.
At both Wistia and Postscript, the number 1 thing we heard from people who signed up and never engaged in a deep way was that they just wanted to see what the product could do, poke around, and see it in action. They weren’t ready to buy or use it in a meaningful way yet. They were just doing some research.
So my teams built interactive demos in-house.
What’s really fascinating is that with a typical PLG offer, say a free trial or free plan, maybe 3 to 6% of your website visitors will sign up. And that’s if your website is well optimized.
When my team put the demo we built ourselves on the site, more than 25% of visitors to those pages engaged in a meaningful way. They got to play with the product and really experience the value.
Total signups went down slightly, but the number of activated accounts went way up, because the people who did sign up were way higher intent. They’d already built momentum in the tool before they ever created an account.
Using your product to market and sell your product is where the magic happens.
Today, you could build this in-house like my teams did, or cobble something together with AI. I’d use a tool like Navattic, which lets you build, ship, and host interactive demos right on your website, and spend my team’s time on other workflows.
Quick disclosure: I’ve been an advisor to the Navattic team for the past few years. I really respect what they’re building and how they’re building it, so please factor in my bias.
Consolidate 3 tools into 1
The second change: I’d consolidate a few of these tools so the data from each step lives in one place.
At Wistia and Postscript, we bought Mixpanel, FullStory, and VWO to do 3 separate jobs: analytics, session replays, and experimentation. Today, some newer all-in-one tools cover all 3.
You’ve seen me mention PostHog 3 times in this piece already. I’ve actually never used it myself, but so many of my clients do that it’d be the first all-in-one tool I’d evaluate if I was hired to lead growth at your company.
How the 6 steps work as a system
Here’s the full system, step by step, with the tool I’d use at each one:
Find the leverage with an analytics tool like Mixpanel.
Figure out why with session recordings in FullStory.
De-risk potential fixes with user testing (UserTesting.com or Userbrain).
Test the new experience with an A/B testing tool like VWO.
Nudge users inside the product with a tool like Pendo.
Follow up outside the product with triggered marketing comms in a tool like HubSpot.
Then I’d add an interactive demo to let the product sell itself (Navattic), and look at consolidating steps 1, 2, and 4 into something like PostHog.
If you’re leading growth in your org and want support breaking through your growth plateaus, running your growth team more effectively, and getting to the next level in your career - that’s exactly what I do as a coach.