How CommandBar Grows: Owning the PLG-as-a-Service Layer

Never forget Clippy, “The fucking clown”

👋 Hey, I’m Jaryd! Here to help you build and grow your product with actionable deep dives into the growth motions, strategic playbooks, products, and founding stories of world-class companies. View all my previous analyses here.

---

Hi, everyone!

There’s a universal truth that as users of countless software products, we can probably all agree with: Popups suck.

And—just like that party host who tries to show you around their new house, pointing out their gallery wall and all the nicknacks you don’t really care about—so do those 20-step product tours. Skip, thanks!

Even as someone who studies products and builds products for a living, I can’t close a popup fast enough.

But up until now, these annoyances have been somewhat of a necessary evil. An evil I’ve guiltily deployed many times across the products I’ve built for my users to probably rage click out of faster than Batman can disappear from a conversation.

This has been especially true for PLG products, where tours and popups are how product builders and growth teams drive adoption.

But, luckily for the world of software, in 2020 James Evans, Richard Freling, and Vinay Ayyala co-founded CommandBar with a key insight: Digital adoption products are a frustration-center because they push information to users.

Like all brilliant ideas, the solution seems so clear when you hear it. CommandBar pulls users toward information when they need it and turns user assistance into a delight-center.

The easiest way to think about what they do is like this: CommandBar allows product teams to embed an intelligent AI agent on top of their software to help their end-users perform actions, fetch data, and co-browse with them to show how the product works.

As James told me in our interview, it’s like having a friendly human assistant standing by your side who anticipates your needs. Copilot is a true successor to the antiquated tree-diagram chatbot— it’s just miles more personalized and helpful.

Through their comprehensive user assistance suite, CommandBar helps their B2B customers with all the vital parts of their own growth motions: Onboarding, activation, upselling, and retention.

Ironically, despite CommandBar being a sales-led company, they’re in the business of helping product-led companies do PLG.

That’s why my thesis is that CommandBar will win because they’re building the PLG-as-a-Service layer for the generation of companies scrambling to be product-led. In a way, like the strategy of selling picks in the gold rush.

And they’re already killing it. They’ve raised ~$25M, have hundreds of customers, are reaching around 25M end users, and are growing at 300% a year.

But, as we’ll see in today’s deep dive, CommandBar has found growth momentum by bucking the common wisdom that startups should create a category.

If you only have a few minutes to spare, here are some tactical things product builders, founders, and senior tech leaders can learn from CommandBar in just 5 minutes.

On growth 🌱

• One of the most impactful, lowest investment, and least risky growth levers is to make what you already have much better. While growth through “new” seems more sexy, you likely have much more juice to squeeze out of your existing product.

• Offering your audience unique and aggregated first-party data from the larger market is an incredibly strong carrot. Since CommandBar is a layer sitting across many PLG platforms, they are able to aggregate insights across many companies and share novel insights with their customers on how they can do PLG better.

• Product-market fit maps very closely to how repeatable your customer flow is. If your demand is fragmented and people want you for different reasons, then you're probably good at sales/marketing but might not have strong PMF.

• When going after your ICP, don’t stop at title. Push further into pain points, company structure, what other tools the company uses, etc. You need to go deeper and ask, “Are these the best folks with this title that we could meet?”.

• The quality of company-generated content is increasingly becoming a signal for the quality of the product and a whole-product differentiator. CommandBar has an insanely good blog—one of the best I’ve seen—and they told me that it’s very hard to make the 1,000-word post work these days. Either content needs to be very deep (their strategy…and mine), or very snackable. The “Content Written For Google” level stuff isn’t engaging, and to stay defensible in the Generative Search era, you have to optimize for people staying on your page.

• Tapping into sponsored podcasts and newsletters as a channel can be an incredibly effective way to reach your market. But, you need your audience and messaging— with a differentiated narrative and counter-positioning— locked in before you do creator-led growth.

• If your product is found through searching, and not discovery, then G2 reviews can be a strong source of leads. When products are evaluated by comparison shopping (like B2B software in a known category), G2 is the place where that happens. You need to actively ask and incentivize your customers to review you.

On strategy 🧩

• Disrupting an existing category is often better than trying to create a brand-new one. At least, to start. James said, “Every founder wants to be an innovator; building a new mousetrap the world has never seen and building a market before any competitors catch on. In most cases, I firmly believe that process is hard mode.” Easy mode is taking the insight behind your mousetrap and using it to disrupt an existing category by adding your own rules to a known game.

• You can always start in an existing category, and then once you’ve found success, pull the market into a new one you create later. Perhaps the best way to create a category is to go through a category.

• It’s possible to be so good that your customers want to show your branding to their customers. This is elite-level trust, and happens when you prioritize for the end user experience even more than your direct customer. Although, both usually win in this case.

• Interoperability is how you can make your product hard to replace. Embedded products win.

On product 💎

• When playing an established game, you can add your own rules by building a narrower version of what your competitors have, with just enough functionality to compete, that still preserves and highlights your core insight. This isn’t just about Tablestakes & Differentiators, but more so the opportunity to differentiate on table stakes and focus more on the outcome that the customer is looking for. You don't have to copy.

• It’s essential to align your AI pricing with your overall product strategy. You need to think about what outcome you’re trying to optimize for, and based on that, determine the risk-reward for capping usage.

• As much as possible, you want to focus on pulling, not pushing. It’s as basic as this: Listening and learning beats talking. Yet, many products don't abide by this fundamental law of communication.

• Another golden rule is this: Don’t annoy users. This starts by thinking about what annoys you as a user, then working backward to make sure you’re not forgetting to empathize with your own product’s users.

• James believes the theme of PLG has destroyed more enterprise value than it has created. To his point, PLG works once you’ve achieved PMF, but before then, being sales-led helps you understand why people are coming to you. Plus, the downside with PLG is that the motivation to talk to you is so low once people have access.

On startup building 🏗️

• Don’t overthink the mechanics of PMF, and dwell on the question: “Do we have it?!”. The focus should be solely on executing, speaking to people, and learning from what you’re doing.  

• The pitch deck is first and foremost for you as the founders, and investors second. “We’ve always been clear-headed that the most valuable asset in our company wasn’t the capital, but our time as founders. We looked at our seed deck as a memo to ourselves of why we - self-proclaimed smart people who could do a bunch of other things with their time - should invest the next decade+ of our blood, sweat, and tears into this company”, says James.

• Spending too much time on the design of your deck is a red flag. It shows you’re either prioritizing the wrong things or “you are trying to hide bad ideas with pretty slides”

• It’s a misnomer to think you’re not a good founder if you’re not doing deep work. James says sometimes the the best thing you can do— at least once you have PMF—is unblocking someone in 3 minutes vs an hour. He has days where he chooses to just do meetings and sit on Slack.

In today’s analysis, we’ll explore how CommandBar is reshaping what product adoption looks like as they seek to own the PLG-as-a-Service layer, as well as their contrarian Trojan Horse strategy for category creation. We’ll cover: 

1. Origins: Cmd+K; The Universal Search Bar. When CommandBar started as a single product company after an initial idea pivot, the founders had a strong opinion on what they wanted to bring to market—Cmd+k-as-a-Service. But despite getting into YC and seeing signs of early PMF, they realized creating a category was deeply overrated.  

2. Repositioning: The Gravity Of Markets. Starting in the late 90s, with Microsoft’s infamous Clippy, product builders have realized the importance of helping users learn and adopt software. That’s resulted in the commoditized Digital Adoption market with minimal innovation. CommandBar’s thesis is that these products suck for users and are ineffective for the same reason Clippy is dead: Pushing info to people is a flawed approach.

3. Product: Command(+K) and Conquer. To unlock growth, stop playing the game on hard mode, and show the market why CommandBar’s approach of assistance through pulling is right, James swallowed a hard pill and repositioned CommandBar to compete in the existing category. Within just 3 years, they’ve built an AI-powered user assistance product suite for product-led SaaS companies and are leaders in the market.

4. Playbook: Become the Gold Standard That End-Users Demand. CommandBar’s move has been to turn product adoption (a frustration) into user assistance (a delight). As a result, they’re following a similar path to products like Stripe and Plaid: Users trust these products so much, that companies want to show off that they use them. This is the ultimate awareness flywheel. And as it spins faster, the big picture is to pull their market once again, back into their own defined category.

Great! Hatch secured. All systems ago. Let’s begin our descent.

---

---

1. Origins: Cmd+K; The Universal Search Bar

Like so many incredible products, the idea for CommandBar came from the founders pulling the thread on an internal product they made for another startup they were working on.

James, Richard, and Vinay were buddies from college building codePost—a tool for computer science teachers to grade and drop feedback on student code.

Despite having no prior startup experience, the trio had a knack for product building. They were great at learning from customers and quickly iterating based on feedback. But, this led to a product with a lot of functionality where the learning curve continued to increase. As James said, this led to an important insight; one that would lead them to CommandBar.

We learned with codePost that shipping fast and being super responsive to user questions and feedback was a competitive advantage. We also learned that when you do that, your product can get complicated and hard to use quickly.
codePost became kind of one-size-fits-all but in a way that meant confusing-for-most.

James Evans (CEO)

Being the savvy builders they are—and hating the idea of intrusive popups and classic onboarding tactics—they came up with a solution to help users navigate around and get the most value out of codePost.

They made a terminal full of contextual shortcuts for their users, allowing them to teleport anywhere on the app without having to explore menus or figure out what to click.

I love how James describes it:

It was a “command bar” that allowed us to remove the entry points to a lot of features: instead of a button leading you to X, we could allow you to search for X. That way, any user could quickly find what they cared about, but the UI would never be cluttered or overwhelming. Basically this was a way to scale a UI but keep it discoverable. It also had the unexpected effect of increasing our product velocity; we could build narrow and experimental features without sacrificing any UX.

A true magic widget, all at the tap of Cmd+K.

It worked incredibly well. So well, that the founders started spending more time thinking about this widget than their actual product.

That's the biggest sign there is. Follow what excites you in the early days, even if you’re technically doing something else.

So they decided to turn their magic widget into a productized Cmd+K-as-a-Service startup, believing that other product builders would be facing the same problem they solved for themselves.

To a universal software problem, they incepted the idea of The Universal Search Bar. 🪄

CommandBar was born in the Spring of 2020, and the team started building cmd+k for any app during the Summer batch of YC.

Their core belief was that users spend too much time learning how to use new interfaces and that a search-first approach was simply better.

The first product gave teams the ability to configure the search bar so the default state could be personalized based on end-user attributes. For example, by user type, level of engagement (i.e. power users vs unactivated ones), and activity history within the product.

This is what the first product looked like, and how they packaged it in their seed deck.

Quick bit of advice I loved from James regarding this first slide of their pitch deck:

My biggest advice (and learning) from iterating on this slide is to be as clear as possible about what you’re building early on in the presentation. If you don’t, everything that follows will just lead to more confusion.

Today that first product, called Spotlight, looks like this. Still very true to the core concept, just 10X more powerful.

Spotlight was, and is, a window into any application. It allows people to see and go deep into a product without having to go anywhere.

At first, James positioned CommandBar as a brand new paradigm for interacting with software. A new category to disrupt the boring, annoying, and commoditized Digital Adoption space.

And rightly so, it was.

But educating the market on something different and selling your product in a newly minted category with no competition is hard for three obvious reasons:

1. Nobody has the budget for your new category

2. New categories are seen as risky, and buyers have a Fear of Fucking Up

3. When you create a category, you remove helpful reference points

Nonetheless, they had real traction. But traction can also be like a Siren’s song—pulling good sailors into the sea if they're not careful.

Signs of promise in the eye of the category creation storm

In the year after YC, their universal search bar was spreading faster than they expected.

They were winning nice early customers and adding great logos like Gusto, LaunchDarkly, ClickUp, and Netlify. All inbound, I might add. This pushed them quickly toward $1M in ARR, and powered them through their Seed and Series A rounds, raising $25M in total.

To many founders, they’d go home singing at this point about finding product-market fit.

But not James. He saw this as a false summit in their journey to PMF.

At this point, he was doing all the selling. And despite closing deals, what he realized was that each deal won felt very different. As he put it, “Each company thought we were something different. And our buyers were very disparate. We had leaders from Product, Engineering, Growth/Marketing, and Customer Support/Success.”

• Some thought CommandBar was a power user shortcut tool

• Some thought CommandBar was a dev-tool-specific alternative to onboarding

• Some thought CommandBar was a way to reduce support tickets

To James, product-market fit maps closely to how repeatable your customer flow is. I’d read that sentence again because that’s golden wisdom.

Does everyone who wants to use your product have the same problems, and get excited about the same features? Or is everyone a bit different and using your product randomly and asking for features that don't really make sense? If the latter, then you're probably good at sales/marketing but might not have strong PMF.

— James, via a Reddit AMA

So, seeing the traction but realizing it was fragmented traction, James and the team made a big positioning decision, based on two additional observations, that led them to broaden their narrow product sooner than they thought. 

1. Their customers wanted more ways to influence their end-users, especially the ones who didn’t have intent yet.

2. The more they studied how influencing users without intent was being solved, the more they saw a huge opportunity in the established category.

CommandBar decided to join the larger market and play the easier game—except, they brought along their own set of rules.

2. Repositioning: The Gravity Of Markets

To understand why CommandBar’s product is so necessary, let’s take a moment and pay our respects to any early internet user’s long-lost frenemy.

Never forget Clippy, “The fucking clown”

In the early ’90s, most people had never touched a computer, and almost half of America had never heard of the Internet. Plus, the small circles of people that were using them were interacting through command-line interfaces.

So, it’s easy to see how the concept of navigating software and using things like menus in a GUI was a foreign concept to many people.

That’s why when Microsoft was pushing their PCs into the home, they knew they’d need a strategy to break down the adoption barrier and make it easy for non-technical people to learn and use their software.

Their first idea to make people more comfortable with computers in their homes was to make navigating a computer like being in your home. They did this with Bob.

Instead of icons, menus, command lines, and lengthy documentation; a user’s computer was modeled to look like a home. It had a front door when you’d sign in, and various rooms around the house (like one for Office) with things you could click on to open up applications (like Word). And importantly, you’d know about all this stuff thanks to some anthropomorphic assistants at the bottom right corner.

The obvious hope was to create a vibe of familiarity.

Except, people hated it, and Bob was shut down in ‘96. But the concept of an onboarding character lived on…

Bob….clicking on the paper would open Word.

…one of them being that know-it-all, googly-eyed, paperclip.

“Would you like help?”, repeated Clippy on loop.

Clippy—which was nicknamed “The fucking clown” internally at Microsoft during the development of Office—was one of the most frustrating user interface components in computer history.

If you were writing a doc, Clippy would chime in.

The idea here was sound. Clippy would theoretically show up during the moments Microsoft thought you’d need an extra hand and walk you toward the relevant action. This was the first real pass at consumer-centric digital adoption. But, Clippy ground people’s gears for a few reasons.

1. He lacked any personalization. Clippy had no idea who you were or what you were trying to get done while using Office. He just made assumptions, and with those assumptions, told you what to do.

2. He didn’t care to learn. Clippy had no user context, but he also never asked you anything to try to gain some. He didn’t know how to listen.

3. He was invasive. Whether you needed him or not, Clippy had a knack for interrupting you. Maybe you didn’t need help, or maybe you were focused and didn’t need help now. Either way, he’d come raising his brows at you.

4. He infantilized the user. Power users comfortable with Office didn’t need an animated paperclip to tell them what to do.

He was eventually snuffed out for his flaws. But, his legacy lives on—nudging people towards action with pushed information is how companies still try to flatten the learning curves of their products.

That’s because the same underlying problem still exists: software products can be hard to learn and use.

Unfortunately, most companies who carry his torch are building on the same premise of how onboarding should be driven.

More Clippy’s…

All you need to know is that the Digital Adoption category is made up of all the products offering in-app tools designed to help with product adoption, feature adoption, and reduce user friction as they learn some new app.

These are the popups, checklists, tours, pulsing hotspots, help widgets, etc you can find across companies like Appcues, Pendo, and Userlane, or as home-grown internal tools.

Sometimes these things can be helpful, but only—and I say this generously— sometimes. They’re better than Clippy, sure, but most of the time we all still try to close them as fast as possible.

Given such a common user behavior, it’s surprising nobody latched onto that problem besides James. As he said when recounting where the idea for CommandBar came from.

The status quo for reaching users in-product was largely interruptive pop-up messages. As users ourselves, we always found these rather annoying and closed them quickly. Turns out, this was increasingly the case for our customers as well – declining engagement rates because of growing user fatigue with pop-ups that they were trained to expect to be unpersonalized and unhelpful.

This slide from CommandBar’s seed deck shows how they viewed the market:

It’s very easy to agree. Their hypothesis was that the most user-centric tools aim to personalize experiences by comprehending user intent and refraining from inundating the user with company-centric information that they may not find relevant.

This means fewer generic popups and more tailored guidance.

CommandBar is our attempt to be a layer on top of all products that can help users by detecting and acting on their intent, using AI.

Fired up with this thesis, CommandBar’s first strategy against the entrenched and crowded market of Digital Adoption was to create their own pond and position themselves as a new category of software. As James said:

At the time I felt like the only startup "worth" starting was something novel that wasn't just an incremental improvement on something that was already in market. Since then, I've come to appreciate the gravity of markets. Selling something great into an existing market is 10x easier—it's easier to find buyers, it's easier to explain what you do, it's easier to justify budget, etc.

Luckily, as we already touched on, James knew to see through the early PMF signals they were finding in this new pond.

He was not going to be the founder who missed the forest for the trees.

⚠️ This post will soon be cut short in email. To keep reading click here. ⚠️

Switching to easy mode

James put it this way, and I couldn’t agree more: Startups are never easy, but you don’t get extra credit for making it even harder.

And building a category is certainly making it harder. When we spoke, he told me that it might be faster to get to $1M in ARR as a category-creating startup, but it’s much harder getting to $5M-$10M a year. That’s because you have to cross the infamous chasm and find your market beyond early adopters.

And there is no harder mode than chasm crossing. Whereas when you find a way to be noticeably and measurably better than the status quo in a market, you skip this level. And the status quo very clearly sucked for both sides: companies struggled to generate more than sugar-high increases in engagement, and annoyed end-users immediately closed the pop-ups.

Who needs to build a new category when the conditions are just perfect for category disruption: unhappy customers for expensive, low-performing, products?

The inflection point for CommandBar came when one prospect said this to them: “If you can replace Pendo for me, I can buy you. If you can’t, it’s going to have to wait a year.”

This time, they leaned into “yes”, and James repositioned CommandBar, hopping them outside of their own little pond into the bigger Digital Adoption one. Cheekily, they rebranded the category they joined to User Assistance—symbolic of how they wanted to compete against the myriad of mediocre products. Basically it was them saying, “Fine, we’ll play your game, but we’re going to play it on our terms”.

Baller.

When we repositioned, we no longer had to explain why what we did was useful. Product teams already know what product tours and surveys are for.
Instead of taking 25 minutes and multiple customer examples to explain what we do (like in category creation hard mode), it now often takes two minutes. “We’re evaluating digital adoption vendors and came across you because of your unique approach.” Great. Qualified. That’s easy mode in action. We even had people adopt our wording and ask specifically for user assistance.

— James

This move doubled their growth rate.

It’s also a move we can all learn something important from: the market always wins.

While making something new and designing a category is the latest trending flavor, it’s not the only way. Sometimes founders need to follow the pull of the market and the path that already exists instead of holding onto their original assumptions.

[Repositioning] made it’s easier to explain what we do to potential buyers, but also people in our lives — family, friends, etc. Along with the signals that this strategy was working, this change made everyone feel better and more proud of what we’re doing. “We’re replacing shitty popups and chatbots” (easy mode) is a lot clearer than “we’re building this search bar thing” (hard mode).

Now that we know what game CommandBar is playing, let’s see the new rules they’ve introduced.

3. Product: Command(+K) and Conquer

After pond hopping—which mechanically involved (1) a marketing & positioning exercise, (2) a new product roadmap to move quickly from one to many products, and (3) an updated GTM motion—it became clear they were finding PMF on the ascent of the true summit.

Here, momentum was picking up and they finally felt the pull of demand. The product-market fit signs they saw were:

• They started being able to reliably drive demand. Blog posts drove leads. Conferences drove leads. Sponsorships on Lenny’s Podcast drove leads. And all the leads looked similar, vs disparate like the first time.

• Their customers started to repeat their unique messaging to them. “I found you on ___ and it seemed like you had a different solution to user onboarding that was more intent-based.” The key was that people would use their own words, like how they rebranded the category to user assistance, during conversations. This proved what they were putting out there was resonating with people.

• The first 10 minutes of every discovery call had a very similar feeling. Despite selling to a variety of personas (Support, Product, Growth, Marketing) and covering a variety of use cases (ticket deflection, activation, engagement), a very similar pattern at the beginning of calls emerged.

Expanding on that last note, James has an excellent point:

Before our positioning change, we were reaching product, growth, marketing, and CX people. After our positioning change, we were still reaching product, growth, marketing, and CX people. BUT, we were reaching people in those departments who (a) had immediate pain that we could solve and (b) knew something about what we were offering. This made a HUGE difference. The repeatable point here I think is that you shouldn’t stop at “we’re reaching folks with the right title”. You need to go deeper and say “are these the best folks with this title that we could meet”.

Same game, new rules

There are many big-picture levers that a company can pull on to grow. For instance, new growth can be charted based on:

• A market penetration strategy. This involves increasing market share for existing products by selling more within the current market.

  • Example: Coca-Cola running more ad campaigns to sell more Coke to current customers.

• A product development strategy. This involves creating new products to sell to existing markets.

  • Example: Apple creating the Apple Watch to sell to their existing customers.

• A diversification strategy. This involves entering a new market with a new product.

  • Example: Amazon moving beyond ecom and into cloud computing with AWS.

• A market expansion strategy: This involves entering new markets with existing products.

  • Example: Starbucks rolling out into new countries

It’s the famous Ansoff Matrix for growth bets:

But, what this chart fails to show is probably one of the most impactful, lowest investment, and least risky growth levers there is: making what you have much better.

This point was made perfectly by

Ken Rudin

in one of my all-time favorite reads of Lenny’s: How to accelerate growth by focusing on the features you already have

In my experience, you can drive significantly more growth by focusing on getting users to engage more with your existing key features and user flows

In other words, if you get your core product and experience wrong, no matter what new features you tack on—you just can’t grow. At least sustainably.

And that’s the growth lever CommandBar puts on a silver platter: their product helps their customers grow by removing end-user frustration, creating delight, and amplifying their existing product.

It’s the same game other Digital Adoption players are trying to play. But, the big experience difference with Commandbar is that by having an AI-powered guide standing behind each user, it’s like having a well-trained-to-your-preferences human butler for everyone without all the cost and overhead of maintaining a live team.

It’s the deepest level of personalization you can get with an onboarding product that I’ve seen, allowing their PLG customers to craft very individualized ramp-up experiences for their users.

Going back to “the fucking clown”, CommandBar’s approach to user assistance took a lot of inspiration from Clippy’s shortcomings. In their autopsy of the paperclip, they came up with a few rules that differentiate them in the market.

These 4 rules are also nice generalized product-building principles.

Rule #1: Pull, don’t push

Unlike Clippy's intrusive approach, CommandBar is all about user-initiated interactions. It’s about surfacing relevant assistance when people need it, rather than having “help” forced upon them in the wrong moments.

Most onboarding tactics push content to users. The “take a tour”, “click here”, or “watch this” prompts. This always means people will see some stuff they don’t care about and will struggle to find some stuff they do care about.

That’s why CommandBar focuses on pull. Pull means giving users some way to tell you what they’re trying to do, then using personalization to give them the lowest friction path to getting that job done.

Listening is a superpower.

The obvious benefit of enabling pull is you make it possible for users to find the stuff they’re interested in without having to read their minds.

Rule #2: Learn from user actions

Because CommandBar uses pull and allows users to specify what they want, they’re getting direct input from users that CommandBar then stores as search logs, allowing product teams to learn from intent.

This becomes very interesting data that you can turn into personalized outreach for upselling, relationship building, or roadmap ideas.

Rule #3: Help users avoid distractions

Clippy would happily interrupt you while you were working. CommandBar is all about minimizing distractions and keeping users in flow.

And the main way they do that is by keeping users inside the product.

This is important because as we know, CommandBar is for PLG companies. And the entire point of this model is self-service. And self-service means people are learning about a product on their own time. And doing something on our own time means we can easily get distracted for a gazillion reasons.

That’s why activating a PLG user on their first session is so important. CommandBar is designed to optimize for this by making sure nobody has to navigate to anything outside of the product or even the page a user is on.

In other words, with CommandBar there are no potential offramps.

Beyond all the content and actions being served in-app, CommandBar goes a step further by pointing people toward the most useful subsection of content that matches their query, whether that’s a sentence in a help doc or a video clip. This is very useful for distraction avoidance and just shows how much they value the end-users time.

Rule #4: Guide rather than interrupt

Alerts just annoy people, plain and simple. Up to 78% of users have reported deleting an app because of notification fatigue. You can probably think of a few apps that overdo their push strategies, which as CommandBar points out is unfortunate since when timed correctly, they can actually be helpful.

Therefore a big part of CommandBar’s suite is tools that intervene at the right moments. This is necessary because you can’t totally rely on users telling you their problems—a platform needs to make smart deductions. For instance, if you’re rage-clicking around or are navigating in a way that the AI picks up isn’t achieving anything and is predictive of confusion, the assistant can offer help: “Were you trying to do X?”, or “You seem confused, see this guide that explains more about Y!”

CommandBar’s answer to “intervening at the right moment” is their AI-guided Nudges product.

With it, CommandBar makes any action nudgeable, where you can tailor a nudge based on user attributes like:

• How new is the user?

• What type of user is this?

• Have they used this page before?

• Does this user’s past behavior (including mouse movements) suggest they are confused?

• On other sites that use CommandBar, does this user tend to like support content?

That last point alludes to a very interesting strategy that we’ll cover when we get to CommandBar’s playbook. 🤐

The product breakdown: The PLG powerup store

The CommandBar platform is like a candy store where PLG companies can go and find ridiculously delicious and helpful product powerups for their growth motions.

Each feature has a clear standalone value, but the more of them you use, the more of a value multiplier effect you’ll be able to give to your end users. Simply, CommandBar is going for the bundled platform play: To be the only tool PLG companies need to drive continuous onboarding.

And the way they’ve bundled this store is very interesting, with all components working together to wash the hands of the others. In short, their product is made up of:

• Nudges, for proactively suggesting experiences for low-intent users.

  • This offering bundles form functions like tours, announcement modals, surveys, banners, and checklist/questlists. These are all the neutralizer features to compete in the existing category and become a replacement for tools like Pendo, except, these are flexible, non-annoying, and crafted with finesse for the end-user experience.

• Copilot, for actively providing experiences that respond to high-intent users asking for help.

  • This is where CommandBar has their truly differentiating features. There’s HelpHub, Spotlight, and by far their most novel feature: Copilot. Copilot is their AI-powered assistant who can help users do almost anything, including answer questions, perform actions on the user’s behalf, as well as co-browse with them to show them what to do. Copilot learns about user preferences (like reading to learn or exploring to learn) and is trained on user and business goals as it develops its assistance strategy per user. It’s also available on mobile.

• A game-changing user model for tailored onboarding.

  • The premise of CommandBar is personalization, and they deliver on that promise with unique segmentation criteria, as well as control features like

    • Rate limits, where you can choose to limit the number of product tours, announcements, and surveys your users will see in a given period.

    • Rage clicks, where you can choose to display a nudge whenever your users seem frustrated and are repeatedly clicking an area on your product. This is one of the many behavioral triggers available.

    • Strike rate tracking, where you can track users who seem annoyed based on how many times they've rapidly closed a nudge. Their analytics then makes it easy to experiment with your nudges so you can improve your strike rate.

• Simple and actionable analytics.

  • The dashboard shows you everything you need to know about how users are interacting with your CommandBar instance. From here, everything is measurable and optimizable.

• Relevant integrations.

  • Obviously you can plug your tools into CommandBar with built-in integrations. No code required. This, along with stuff like admin controls and security, are their tablestakes—the basic expectations for an enterprise-ready SaaS product.

How their AI pricing strategy supports this

Figuring out how to price an AI product is tricky. There are two camps really:

• Usage-based pricing

• Value-based pricing

Rather than directly tying their AI costs to usage, CommandBar's approach leans towards value-based pricing. They correctly recognize that the true value of their AI features lies in their availability. This strategy contrasts with a usage-based model, which very much like having an hourly lawyer, can create uncertainty for customers and potentially misalign incentives.

CommandBar folds their AI pricing into their flat-rate subscriptions. It’s one and the same. And that’s essential because CommandBar has two things that must be true for them to execute their big picture strategy (which we’re about to unpack):

1. CommandBar must be consistent: For Copilot to help, it needs to become the go-to for users. Building that trust requires consistently answering user questions so well it resolves the issue they came for.

• CommandBar must be ubiquitous: Using Copilot won’t become a habit for users if it’s only available on certain pages or for certain topics. Only if Copilot is ubiquitous across their customer’s products will it become their default help solution.

These are the insights our pricing flows from: If we shut down Copilot because our customer ran out of AI credits, it’s no longer consistent. And if customers worry about API usage, they might limit Copilot to the hardest to understand parts of the product. That’s why we landed on a simple subscription: It’s the easiest pricing model to provide Copilot’s value.

— James, via How we priced our AI SaaS product (and the mistakes we avoided)

Plus, by bundling their AI features within existing pricing tiers, CommandBar reinforces the idea to customers that Copilot is an integral part of their product's value, not a separate add-on. Their killer differentiator is never limited.

Of course, there’s the risk of abuse and losing money when you’re charged variably but you’re offloading that cost to users at a fixed price. CommandBar clearly believes the benefits of value-based is greater than the risks of usage-based, but they do have two safeguards in place to prevent catastrophic anomalies.

1. Rate limiting based on Monthly Active Users (MAUs)

2. Monitoring usage to identify and address unusual surges

Something to noodle on when considering how to charge for an AI product.

4. Playbook: Become the Gold Standard That End-Users Demand

In this final section of today’s deep dive, let’s talk about strategy. These are the 5 things that really stood out to me in my research.

1. Their trust strategy

2. Their interoperability strategy

3. Their hard-to-replace strategy

4. Their Trojan Horse strategy

5. Their sales-led strategy

Let’s review.

What Stripe and Plaid taught us: It’s possible to be so good that your customers want to show your branding to their customers

The moment they’re given the chance, most companies hide the branding of any vendors they’re using for end-user-facing products. Beyond that, many companies will only choose software if that’s an option.

It makes sense—you just want your branding on your app. Why have another company’s logo embedded on your site, especially a company that you’re paying?

Of course, as the providing company, it’s wonderful having your little “Powered by Us!” copy being flaunted around the internet. You get to enjoy the embedded flywheel, where your product is marketed to your customers’ network of users.

Example: If CommandBar sells to 1 company, but that company has 10,000 customers, CommandBar’s effective reach is 10,000. That’s free broadcast advertising.

Unfortunately for most of the products that have this type of amplified reach, their branding gets disabled, limiting the power of the embedded flywheel.

CommandBar is in the same camp, given users on higher-tier plans can remove the branding. The question becomes, can CommandBar show their customers it’s more valuable to keep “Powered by CommandBar” on?

I believe they can.

Very few companies have managed to do it, but Stripe and Plaid are top-of-mind examples. When you see Stripe or Plaid, you trust a product more. I personally don’t want to use any financial product that requires account connections if it’s not done by Plaid. And I’m far more likely to buy some product if it runs on Stripe because I trust Stripe.

Stripe and Plaid have created specific layers in the financial ecosystem that are often more trusted than the products using them.

CommandBar, as the onboarding layer for many new SaaS products, has a rare opportunity to enjoy something similar.

What CommandBar needs to sell to their customers is the power of this familiarity effect. The pitch is simple: When people see CommandBar on your site, they’ll have more trust as they know how to get around and get things done.

When you achieve something with CommandBar, it feels good because the time between intent and action is short. So you want to use it again. And not using it is like ordering something on Amazon and not getting free-2-day shipping: it just feels subpar.

— James

So, just like how walking into a party with a bunch of strangers is infinitely more comfortable when you arrive with a friend, CommandBar can unlock something similar by allowing users to try a new product with their AI assistant.

Next, let’s look at one thing that will make this sell much easier.

The interoperability of a user’s preferences

Aah, interoperability!

If you don’t look at this picture and immediately hear Mark preaching why “Interoperability” is the future of the metaverse, then you may have been under a rock circa the time when Facebook changed their name.

Anyway…what is interoperability anyway?

In short, it means that one product or system can work with another product or system for information exchange without a user having to do anything.

Remember what I pointed out earlier…

On other sites that use CommandBar, does this user tend to like support content?

That is very revealing of a brilliant strategy in the works.

CommandBar’s killer pitch compared to the rest of the market is that each user has a personalized AI assistant who learns about them and can help them get stuff done.

But how annoying would it be to have an assistant that forgets everything about you when you go somewhere else? Very!

That’s why CommandBar is building to be interoperable across applications. If you use CommandBar on Jira, and then sign up to Amplitude, your Copilot will be at the ready, knowing things like how you enjoy tours vs reading, or that you’re a user who often needs to set up admin roles. This just feeds into why end-users will trust new companies more when seeing CommandBar—they know their butler is there.

This is brilliant, and it means CommandBar can sell compounded value to new customers because they can essentially say: “Don’t worry, even if you don’t know anything about your users yet, we do.”

Any PLG company wanting to help their users find value ASAP will see the insane value of this.

This is also a virtuous cycle: The more users rely on CommandBar, the more search and usage data they feed it. This creates a better CommandBar experience, which means they rely on CommandBar more. Repeat. 🤝

Be hard to replace

Once a customer has embedded and is using CommandBar, it’s jarring for their users to take it away. It’s a layer that can be styled to feel native, and end-users who are accustomed to having a command palette and assistant could easily be annoyed having their trusted sidekick removed.

James put it nicely:

I often joke with our customers that turning off CommandBar would be like asking their users to stop using Google search and go back to Yahoo Directory. It just makes the old building blocks feel old.

CommandBar obviously delivers a ton of value to their direct customers, but there’s a special type of lock-in they get to enjoy because of how much value they deliver to the end user.

It’s almost an insurance policy on churn. Similarly to how no company that values their designers would ever cancel Figma because all the designers would leave. 🤷

Power to the end-user.

Their Trojan Horse of category design

We know CommandBar started by trying to create a category and then switched to easy mode—wedging themselves in an existing category.

But, as Fat Joe iconically says: Yesterday’s price is not today’s price! ✊

What that means in our context, is that CommandBar is creating a ton of value right now in this existing market. They’re finding PMF. They’re building brand loyalty. They’re becoming hard to replace.

What was hard mode yesterday is not necessarily hard mode tomorrow. Tomorrow, CommandBar will have a lot more market power and cachet—cachet they can use as leverage to pull the market into a new category.

In short, I think their arc looks like this

Pull → Push → Pull again

It’s super smart, and a great example of James playing the long game. Because when they eventually do pull their market into a new one, CommandBar will get to reap the benefits of being Category King.

The power of being sales-led (and two tips from how they do sales)

The latest wisdom (at least for the past few years) is that if you’re not running a product-led motion, you’re missing out. All the snazziest companies are doing it, so why not run the same growth strategy?

The playbook, as we’ve seen across many companies covered so far, is to start with PLG and then layer on product-led sales.

CommandBar initially thought they’d follow the same path as the Notions of the world. But they ended up deciding to go against the grain, and James told me it was one of their best decisions.

Outside of the hard numbers like average. deal size increasing by 34% and their conversion rate going up by 17%, being fully sales-led allowed the team to:

1. Speak to far more customers (you get to learn and sell at the same time)

2. Have much more successful onboarding (ironic, I know, for a company selling onboarding automation)

Sales brings a high-touch approach to working with customers, and before you have PMF, James believes it’s a huge mistake to not have as many interactions with your customers at the beginning of their journey with your product as possible.

CommandBar runs a sales motion for small startups all the way to enterprise. Nobody gets self-service. What’s more, CommandBar doesn’t even list their pricing publically…every call to action is to Book a Demo.

Of course, they tested this and found it weeded out low-intent users from their pipeline. It also means the customers they acquire always activate and find value, meaning they’re prioritizing building a very strong customer base who have very low rates of canceling.

Two actionable things stood out to me around how they sell.

1. Every demo is personalized to the prospect. In the early days when James ran sales, he’d spend about an hour learning about each company and creating a personal Loom video showing how CommandBar would work for them. This is still a key part of their sales process.

2. They built a Chrome extension that visually shows CommandBar working on potential customers' own websites. This makes their product tangible to people, vs theoretical. Just like how people love hearing their own names more than any other sound, companies love seeing their own products being used as examples vs some “demo store”. Show, Don’t Tell.

For the curious, here’s their sales stack.

On the topic of sales, here’s one last bit of advice from their GTM leader, Joe Skupinsky, shared by Finn Lobsien:

He also told me about mistakes many sales leaders in startups make. He cited two main failure modes.

• First was not having a process. Startups are flexible, but that doesn’t mean you can just wing it. Without a reliable process you’re wasting time by starting from scratch each time. You also lack the ability to diagnose problems: If prospects constantly drop off during the trial stage, you know where to improve. Without process, you’re in the dark.

• The second failure mode was not preparing for meetings. Even if you’re not a founder, customers need to trust you. As Joe told me: “People will be averse to buying from someone who doesn’t have their shit together. People want to buy from people they feel confident about.”

You can find more details on how they sell here.

I’ll leave you with one parting thought. It’s rare to see a company try their hand at category creation, find success with it, yet still push the rudder in another direction to rather go head-to-head in a market where the players are fairly entrenched.

But the way CommandBar has done it—by bringing their own set of rules and POV to the game—they’ve brought exciting new blood to a stagnant category. That’s a GTM strategy more people should give consideration to, because as James minced no words in reinforcing with me—category creation is very hard.

Lastly, if you’re a product builder, growth operator, or founder running a PLG motion, you likely see the immense value potential CommandBar brings to the table. They’re building the onboarding layer I expect to see more and more PLG companies embedding above their products, and you can get in early :)

Schedule a demo. See how it could work for you. Onboarding in the best way.

A huge thank you to James Evans and Finn Lobsien for making the time to speak with me, and to the CommandBar team for writing one helluva company blog. Without which, this piece would have been far harder to write.

I’ll see you all next time for another voyage down deep.

—Jaryd

---

---