Skip to content Skip to sidebar Skip to footer

Max Lytvyn is Correcting the World’s Writing with Grammarly

Max Lytvyn, the co-founder of Grammarly, spoke to us about how the Grammarly team bootstrapped their company to hundreds of employees and millions of daily active users.

Max Lytvyn is the founder of Grammarly, a freemium grammar writing app with nearly 7 million daily users. Having profitably bootstrapped the company to hundreds of employees, Grammarly recently raised a $110m round of funding. In his talk, “Build Before You Sell”, Max talks about how the Grammarly team quietly bootstrapped their product to unicorn-level, by finding the right homogeneous segment of the market.


Hi, everyone. I’m Max and I’m a co-founder of Grammarly. Today I’m going to talk about bootstrapping. I’m going to use Grammarly a lot as an example today. Let me just say a few words about the company to those who are not familiar with it. Grammarly is providing AI powered products that help people improve their communication, help them make their emails, social media messages, documents more effective, clear, and mistake-free.

Grammarly has a very big mission. Improving lives by improving communication. It’s so big, it almost sounds undoable, not just not bootstrappable. Today I’m going to tell you how it is actually possible to do, how this can be done, and we’re doing it now. Millions of users are using Grammarly and it materially improves their lives. I promise you, if you install Grammarly after this conference, your life is going to get better, even if just a tiny little bit, it’s going to get better. This mission is real. We live by it, but it’s so big, it may seem it’s not something that can be bootstrapped, but we did it. Today I’m going to share with you a framework that helped us do it.

Just to illustrate what I mean by good scale by we did it, this is not to brag or there’s just some numbers to show you what I mean by the scale of our company. It all started in 2004 when my co-founder and I decided that it’s time to get real, it’s time to build a real company. We read a lot of manuals, mostly from Ikea, and we assembled our first office in Toronto. As you can see on this picture, my co-founder, Alex, was the brain of the operation and I was the muscle. That’s why he got the window seat. That’s how real bootstrapping looks like. It’s not very glamorous. I admit it, but it’s just a start.

Image result for grammarly cofounders

Grammarly cofounders. Source: BIV

Until 2009, we were focused on plagiarism detection, which is a very niche area, but it’s related to communication. In 2009 our understanding of communication evolved, our vision evolved, and we had new mission. How people communicate better, and we realized it was very ambitious. It was very difficult from a technology standpoint and it had massive market was diverse because communication problems and needs are different for everybody. What do we do? Is it too much to bootstrap? But we decided to give it a try. We had some experience, we had some learnings from previous projects, so we decided to give it a try. What helped us get through it, what helped us do this was a three element framework that I’m going to share with you today and I very much hope it’s going to help at least some of you to do something big and do something impactful.

Step number one may look familiar to some of you. I guess many here know the concept of MVP. If you don’t, Google it. It’s very useful concept. It’s just great concept. But the challenge with that is that most people apply it incorrectly. They just go straight to the product. They minimize their product. They go with minimal, and it’s easy to make this mistake because it says minimal viable product, right? They go straight to the product. You have to start your MVP with your market. You have to pick a small homogeneous segment with one pain that’s solved by small fraction of functionality that you’re envisioning building eventually. This functionality is something that’s going to be useful to the rest of your market at some point.

Let me illustrate it with a diagram, what it actually means, how that works. Let’s say on the vertical axis you have product functionality, product value, how much value you’re creating for users. On the horizontal one, you have market breadth, like the size of your market. The area of the chart is the amount of work you have to do to get there. The blue shaded area is your ultimate vision. It’s 10, 15, 20 years of work. You don’t want to start there, you want to do something smaller, trim it.

What most people end up with is what I call typical MVP is that gray area at the bottom here. It’s broad because you’re building for everyone and it’s very thin because, well, you have limited resources. What you end up with, you’re spending a ton of work to build something that barely qualifies as a toy and people forget about it in few weeks. But if you MVP your market first, if you find this very small group, very homogeneous group, and build your MVP specifically for them, just for them, you’re going to end up doing much, much less work and the product is going to be better so that it not only gets elevated to the level of tool rather than toy, but you can also charge money for it, which is extremely important for a bootstrap product. You get early revenue very quickly.

Just using Grammarly as an example, again, our MVP market was academics, students, researchers, instructors, and that strong pain that we identify for them was making documents, making papers, compliant with rigorous academic standards. That pain could be sold by a subset of checks of corrections, suggestions, that would be needed for other people as well. We were not deviating from our longterm vision of just building a tool that with how people communicate.

Image result for grammarly academic

Just let me give you an example of what this optimization of building that MVP for that tiny small market gave us. For example, pretty much everyone, 99% of our potential market wanted feedback in real time. You type and you see feedback, you see suggestions. But academics didn’t. They were fine with waiting 5 minutes, 10 minutes, and 15 minutes to get feedback on their document. Why? Because they spent weeks writing, so extra five minutes doesn’t matter. They just go get a cup of coffee. That realization, optimizing for that allowed us to not worry about speed and worry about accuracy more, but it also allowed us to cut engineering costs by 50% and cut server costs by more than 5 times and that was just one optimization. Don’t underestimate the step. Try to go with narrow market first. MVP your market first before MVP’ing your product.

Now moving on to step number two, sell before you build. It’s easier to explain with an example how we did it. As soon as we had specifications for our MVP, we went to a bunch of conferences, trade shows for our target market, academics, and tried to sell the product just using bare demos and it was invaluable. It got us feedback on the product, it got us validation of some of our ideas. The bonus point, it also lined up first buyers so that we had first revenue literally days after we release the product, literally days. That’s extremely important for a startup. Again, because basically our first buyers funded the rest of the company. This approach applies not only to B2B, not only to sales. If you are working on a consumer product, it still applies. There’s still ways to do it.

For example, for our consumer-focused features, we used AdWords to advertise upcoming big features and see how people respond to them. If people click on ads in what numbers to help us prioritize those features. Essentially we pre-sold tested features before they were developed with the real users. Here’s an example of a mailing we did to potential educational customers, and this example reminds me of one very important point I wanted to make. When you do pre-selling or selling something that you don’t have like smoke testing, it’s very easy to not be completely ethical. Please don’t do that. It’s possible to do that in a completely ethical way without any bait and switch or anything like that. Be ethical, and just if you put your mind to it, you can completely do it in a very positive and ethical way.

An example of one of the tricks we use to make sure that it’s completely fine, when we sold somebody on the feature that we didn’t have, we actually gave them a free trial that ended after we released that feature. Before they actually paid us money, they saw the complete product that was promised to them. As a result of our approach to this approach to this pre-selling, every customer we sold product that we didn’t have eventually was completely happy and we used them as references to build more and more customers. Be ethical. It’s worth it. It pays off.

Expand product in the market in iterations. That’s the final step and that’s where most people fail. That’s where big dreams die, so pay attention. The trick here, the main principle here, is build for your next segment, not for the segment, not as much for the segment you have now. Just using example to illustrate it, not quite an example, but analogy. What is more valuable, building for 10,000 users that you have or for 1 million users that you’re about to have? Of course the latter. Think about it this way, it is very helpful.

But of course, it’s never black and white. Whatever you build helps both your existing users and your new users. How do you balance that? I use this formula where you basically take the size of your own current segment and the benefit to it and the size of the new segment and benefit to it and factor in the risk. It spits out the relative value of anything new you build and it’s very helpful for evaluating new opportunities and things you can build. This is how this looks when you put it back on the graph which we saw earlier. You expand your market, you expand your product functionality for that market, and gradually you delight more and more users.

Some do’s and don’ts to sum it up and to make this model even more practical and useful. First, don’t bite off too much. Small, quick steps are less risky. They are generating revenue quicker. You’re going to be releasing new features, expanding your market in smaller but quicker steps and each new improvement is going to help you get more revenue, which is very important for a bootstrap company. Then don’t get stuck chasing your tail. This is a big mistake that many people make and let me explain what I mean by this.

Let’s say you had success on a specific segment of the market and as a result of that, all the feedback that you got, all the data that you’re collecting, is going to be about that segment, but it’s not going to teach you about users you don’t have. Users that you have are very poor at teaching you about users that you don’t have, so you have to realize that. You have to break out of it. You have to go out of your existing user base, go out of the data that’s easily available to you and look broader. Look to your next segment, look to you to expansion and keep focused on your vision.

Then finally expect each next step to be harder. Initially, if you’ve done a good job creating your MVP, specking out your MVP, you collected all low hanging fruit, you skimmed all the cream, you killed flocks of birds with one stone, you’d done all that. But once you’ve done all that, it’s over. It’s no longer that easy. Each next step is going to be progressively more difficult. Again, it seems obvious, but most people don’t think like that. They fall into the trap of thinking, “Well, I’m getting better at what I do, I’m getting more resources. My company is growing bigger, so each next steps should be easier.” No, it’s the other way around. If you don’t use your brain, you have done all the easy things in the beginning and each next thing is going to be more and more difficult.

Yeah, that’s it. It’s a very simple framework. I really hope it’s going to help some of you guys build something new. It doesn’t even have to be a startup. It may be a product within your existing company. It can be applied pretty broadly, and it worked pretty well for Grammarly. We started, as I mentioned, we started as a very narrow tool for a very narrow market on purpose. Then right now we have a broad suite of AI-powered products that help people improve different aspects of their communication from effectiveness, tone, correctness, clarity, pretty much all across the board. That’s just the beginning. Thanks.

Comments