How to Kill a Unicorn
chrisfrantz.com
chrisfrantz.com
I can't remember where I read it, but I read a great article that argued that Google Docs was successfully not because it duplicated Microsoft Word (many had tried and failed with this approach), but because it focused on how work had changed and the need for collaboration.
Canva is very, very successful, but I don't think that's at Adobe's expense. By any measure, Adobe is probably one of the best examples of a company shifting to SaaS: https://producthabits.com/adobe-95-billion-saas-company/. Also, regardless of whether they offered their products in a SaaS model (Adobe introduced them in 2012 I believe), there was still free trial.
Maybe a better explanation is that Canva tapped into a different audience - people that needed to produce graphics but did not have design capabilities.
On Figma, it seems to me to be a better online collaboration design tool than what has been offered by other companies. The author seems to forget that companies like InDesign existed at the same time, offered a collaborative tool, and had a similar freeium offer.
I won't belabor the point - freeium might have helped these companies succeed, but I think it is a big leap to say freeium = unicorn killer.
Would you use Google docs if it cost $200 a year?
The ease with which stakeholders, designers and developers alike can work together, either by desktop app or web, is great. Not perfect, but great.
Figma offers design, prototyping, developer handover, team collaboration, a.o. capabilities. The basic service is free and the paid tier is fairly priced. All said capabilities are also integrated so you don't have to use plugins or switch between apps. When working with Sketch, I use Sketch for design, Marvel/InVision/Framer for prototyping, InVision/Zeplin for developer handover, and InVision Freehand/Miro for collaboration. This means dealing with multiple licenses, plugins, and copy-pasting assets.
I also worked with a mid-sized company that was paying around EUR 20,000 just for their InVision license (developer inspect is expensive). With Figma you get most of these services (good enough) for a single price.
So money certainly plays a role but Figma's success goes beyond just the pricing model.
It helps everyone in the company since we had marketing usually being on windows, developers usually being on linux and designers usually being on mac. With Figma anyone could browse the designs and before that (when using Sketch), we had to export designs from Sketch to something like Zeplin just in order for others to view the updated designs. Then designers need to remember to update the exports every time they changed anything.
With the introduction of Figma, we didn't just get the ability for everyone to browse the designs directly, but also make smaller fixes directly in the design together with the designers, remotely or not.
I'm pretty sure that this was the killer feature.
This is a combination of no OS dependence, no installation required, and no license key software. The software being any good is just gravy on top.
Anybody who has ever dealt with IT understands just how much pain this saves them.
But here's the million dollar question for someone making those products. How do you convince your target audience? With the old business models, you'd spend a few million on advertising and hope for the best. If you have a generous free plan, you'll get people signing up, and if you're lucky they'll use it. Of course you need a good product. A good product is not enough in most cases.
I haven't heard of any of the names in the article, with the exception of Adobe. So the author may be talking about some niche where that Canva is actually better than <whatever Adobe sells this week>.
Possibly some web "content" and marketing niche, considering his vocabulary?
You could replicate everything it does trivially with most basic features of Photoshop, or even Paint.NET - its main selling point is that it comes with lots of templates and images you can use with one click, and has a very gentle learning curve (because there isn't much to learn there).
Not harping on you, but this is “Dropbox is just CVS and ssh-copy, trivial” line of thinking.
I have Paint.Net AND a complete Adobe CC subscription, but opt to use Canva for almost all my lighter editing. The selling point is nice looking templates that are context-aware - if you want to pull in more element they will match what’s on screen.
I’m usually done faster in Canva than loading Photoshop would’ve taken
Maybe they are better products, I don't know, but how many people used it because it was free, then realized it was better? How many of those wouldn't have used it if they didn't have a free offering?
Reading them both, it seems the other one is written by a developer and this one is written by a marketing person. The marketing person probably knows more about this then a developer would, but it is fun to see their perspectives. The developer doesn't want to deal with free users demanding resources and the marketer wants to position the product to be profitable.
Personally I trust developers more than marketing people, and I generally have a bit of disdain for marketing folks. They market their trade as exposing the right product to the right audience, but usually what it amounts to in substance is manipulating people. Still, the marketing person's perspective on this particular topic is important to consider.
If I was buying some software for work that costs (say) $500 the process involves many steps, including convincing myself I want it and concincing my boss to pay for it (usually.)
It could take days to give a product a fair evaluation but I have to consider the probability is (say) 30% that I get it in the end, so that the invested time might be "wasted". Say I spend 2 days on eval and I get paid $50 an hour, that is $800 worth of time which I could spend on project work, or 'sharpening the saw' in some other way. If I have to look at four products to buy one I am spending $2400 of time to buy a $500 product.
So I think of free trials as 'free as in puppy'.
If that is the case, then anything which makes it easier to evaluate will make it easier to purchase. If you have to choose between a free trial and spending $500 for a month to evaluate, won't you pick the free trial, even if you are spending $800 of your time to evaluate?
The context I'm coming from is that my current employer has a full featured, forever free offering aimed squarely at developers: an auth/user management system (link in bio if you are curious). The paid offerings have better support and premium features.
Our thesis is that for a focused developer tool (rather than a tool that can be used by non technical users, which are many of the examples in the original post) you need to lower the barrier to entry. Integrating with an auth system is more of a commitment than picking a different tool to create images or manage an email list. It's like picking a database.
I think the original article did a great job of talking about where freemium works (big market, virality, etc), but didn't cover all the places where freemium works, in particular dev tools.
The dev focused tools, a la the indie hacker type, are usually after some very niche user. In this case, you should just make the user pay. There’s not many of them anyways and you can likely command a higher prices because you’re essentially solving a problem that would require them to solve (expensive) or just keep status quo in place (eg manual process).
This is how I like to think about freemium anyways. If product has mass appeal, eg bookkeeping, then freemium makes sense. If it’s bespoke compliance software tailored just for veterinary hospitals then everyone pays
I hope someday you get to know people who enjoy good marketing like I imagine you enjoy good software, so you can change this view. Imagine a different POV: developers sell their trade as solving real world problems, but usually what it amounts to is overcomplicating things or trampling domain experts opinion in favor of bad assumptions.
As in all professions, most work you'll see around is not work to be proud of. Don't use random sampling to form an opinion about the value of a profession.
First, there's pressure on a lot of marketers to sell more this month or this quarter.
Second, there's an expectation to prove that an idea is good using data and to prove the idea worked using data.
Third, there's very little interest in investing in the art side of the trade.
When taken together, you get what we've got.
What am I trying to say?
I guess just that in my experience it's always been the owners and top level managers pushing for short-term, unethical, bad ux decisions and focusing on manipulating instead of serving people.
And it sucks.
I think a lot of us got into marketing because we're creative and technical and wanting to make businesses and products better for people, both consumers and society at large.
Or maybe that's just me.
Edit: the worst part of it is the demoralizing realization that the owners are often right. Most people actually do want McDonalds, like shopping at Walmart/Amazon, and enjoy Facebook. So I guess the worst part about marketing, the part nobody talks much about, is that giving the majority of people what they want usually means low quality at the cheapest price.
Still, I look at a company like burger king, or a toy company, or Disney, or a radio station, and the entirety of them, including the product itself, is marketing. If these companies could take your money without giving you anything at all they would. The idea of marketing is supposed to be to put your product in front of it's users for you. Instead with countless examples you see companies rebuilding their entire product line as a marketing strategy. There is no product anymore, just that last mile of marketing to get your money in their hand. This is so common nowadays.
When the company has lots of money to spend on supporting those free users, having the free plan probably makes a lot of sense to grow faster.
When the company is just you in your spare bedroom, there are fewer cycles to spare on the prototypically-needy free customers, so optimizing for fewer -- but paying -- users can be a good trade.
What I'm trying to say it, there's no magic formula - you need to understand as well as you can your user base, see how long it usually takes them to understand that you do provide value with the product and importantly check out what your competitors are offering in terms of free plans.
—————
At Notion, we used to have a limited free plan where we only allowed users to create 1000 blocks (pages/paragraphs/embeds/checkboxes/database rows) before we directed them to upgrade to the "Personal Pro" plan. Supporting free users dealing with the block limit was a major headache, just as described in the OP, but more importantly the Personal Pro plan was kind of a dead end - few users converted to Pro, and those that did weren't worth much, since the Pro plan is pretty cheap and doesn't have much network effect.
We took a hard look at that free personal plan and how it worked with our goals. Instead of phasing it out, we decided to make it unlimited - now personal users can make as many blocks as they want [https://www.notion.so/personal]. But importantly, we also switched up our segmentation at the same time: we restricted the personal plans' collaboration model and created a new "Team Trial" plan, with the same 1000 block limit, but with all the teamwork features included. Another thing to note about our Team Trail is that it's not time-limited at all, teams can keep trialing Notion until it has clear value to them. We have many fewer users frustrated by the block limit, but those who are are much more valuable, and are getting a lot more out of Notion -- when they upgrade, it's to a Team plan with strong network effects and a path to future expansion.
And all those (now) very happy users on the free plan? They're like jonpurdy [https://news.ycombinator.com/item?id=26059666] - they'll take Notion into their (future) teams and workplaces, even though in the shorter term they're a strain on our systems.
My advice is - there's more nuance to this than "MAKE IT FREE" vs "DON'T MAKE IT FREE". Think about how you structure upgrade incentives around value, and about how your free plans fit into your overall market strategy.
Freemium needs to cover trialing up until the point of an adoption decision. However, the adoption decision isn't about continuing to use the tool. We're in a 'prosumer' world, so the purchasing decision flip is instead about when a user both operationally and mentally phases from b2c -> b2b, meaning changes both in how they're using it and the value they're noticing.
Ex: In Slack's case, it is a team liking it enough that they start having substantive conversations, and when prompted, view paying for a search history to be more than reasonable for a work environment. If they aren't getting commercial value from those comms, they can keep using for free until they do.
Ex: In ^^^ Notion's case, a 1000 block limit may not signify anyone cares beyond entertainment value. But as soon as you want to do secure sharing with colleagues for a lot of work, you're probably doing valuable stuff, and are fine paying as long as reasonable.
Ex: Twilio is ~free for developers as they prove value, but org pays when it gets real.
So it is less about trialing -> addicted, and more about trialing -> committed & appreciates generated work value. The lazy version of this is "addicted individual -> addicted team", so I expect pricing models to roughly follow that. Sort of like a modern version of "Enterprise means SSO so charge more for that", except now more closely aligned with utilization... which is a good thing.
I've found this to especially be the case with new hires who are fresh out of college or graduate school. They have experience setting up platforms for the group projects they do in school and they can often work pretty well for a small team if implemented with intention.
It's not actually so different from Adobe's strategy of somewhat turning a blind-eye to piracy from college students knowing that when they graduate into being design professionals they'll have business expenses they're willing to splurge with.
free for public / oss projects, pay for private
- most personal projs and oss are covered: like slack, that supports successful viral collab!
- but as soon as you are doing commercial work and want that privacy (eg, you are being paid to do it), ok to charge a bit
they eventually got big enough that they could wait even longer to charge. squeezing a few $/mo from tiny accounts doesn't matter if that causes churn before bigger adoption. for a co that straddles both smb + enterprise (1% of customers are 50% of revenue?), that is a suggestive trend for prosumer freemium
another good article here is Greylock's Unit of Value essay
It was my feeling that Figma lacked some of the finess in terms of finer vector drawing and advanced features that Illustrator or Sketch provided but was "good enough" in 99% of use-cases that it didn't realistically matter to their target customers. There are some parallels to the Google Docs X Word relationship I think.
Edit: I read your comment wrong and removed part of mine to make sense. We're basically in agreement.
When it comes to B2C, prescriptions don't apply uniformly, or for that matter scale. Beware.
My only suggestion would be that be honest with the value proposition of your product. Know that your electron-based note-taking tool is not the same as the entire g-suite or ms office.
And finally, remember that no one is more emotionally invested in your product than you are. So let the suggestions trade.
> How Freemium really works [...]
> In reality, there are a number of different ingredients in the recipe you need to nail before you can actually execute a freemium strategy effectively. You don’t need a team of hundreds of people, but it does help to lay out the strategy ahead of time.
> Let's break out the different factors involved in setting up a freemium plan.
- Pricing that scales with usage
- Shareable content
- In-app upsells
- Massive market opportunity
- Pricing innovation on the primary value
I run a freeium Chrome extension (https://www.checkbot.io/) for checking website SEO, speed and security factors where I have the free plan without sign up for multiple reasons 1) low friction so users can try it out first 2) users may be more likely to share and recommend it for quickly checking website issues 3) there's free tools in this space so it may be harder to ask for payment upfront. I'm generally not finding the cost of the free users high.
For another product though, all these factors might be weighted differently which would tip the balance on a free tier making sense. It depends on too many things e.g. marketing budget, competitors (how entrenched they are, how much they offer for free), how much a user needs to try out your product before they realise they need it, differences between free and paid plan, if your product appeals to teams or individuals, support costs.
When your product however solves a very unique problem and is the only product that does that, you don't really need to do such a thing, because the why is self explainatory. E.g. you wrote a format converter that allows users to transfer basic audio projects between all kind of proprietary DAWs. This is so self explaining that only people who have a pressing need to solve that problem will pay for it anyways. And because there is not too much competition in that market you don't have to demonstrate why your solution is slightly better than $trustedopensourceprogram
Ah boy do I wish reading an article like this few years back. This point is exactly what we missed when offering a Freemium plan to users. It doesn't work for products that, by design, are built only for a specific set of users.
> First, if you are missing even one of these, it might not be the pricing strategy for you
Yup. Freemium is for you only if you can to tick _ALL_ of those boxes.
Exactly, freemium does not make sense if you have a limited niche of potential users.
That said, there are strong competitors to Google Analytics, that's the one take I don't agree with.
With Portabella (https://portabella.io) we've gone the restricted free tier route. As a bootstrapped company we have to make that tradeoff between wanting users but also safe guarding against abuse.
For a GA alternative there are many and what you should look for is an analytics provider (if you don't want to self host) that rotates salts. Basically user ip hashed with a random salt that changes every 24 hours is good enough to protect user privacy while allowing you to not count every page view as a unique session.
Of course everyone is counted as a new visitor after 24 hours.
On the rate limiting side, great question. I'm currently writing a series of blog posts around creating your own privacy friendly products (Part 1 is targeted at apps, part 2 will be SaaS) where I'll cover this in detail.
Here's a few key takeaways, if you can provide something without collecting any user data, do it. However for anything involving payment you need to at least know the users email (leaving aside cryptocurrency, I'm a big fan but it's difficult to accept as a bootstrapped business). Additionally rate limiting based on usage (which you therefore need to track) is kind of a given. The best we can do here is make sure user data is encrypted.
So I guess the solution is to have a way to have your customer test your product in a thorough way.
Canva seems all about taking an existing template and swapping in some new photos and text. Of course, this is very useful and I’m sure there’s a huge market for such functionality when creating Instagram posts, etc.
But it’s very different from full-featured image editing (photoshop) or authoring complex vector graphics (illustrator).
Then, you might see the unicorn out there in the mist... the UX person who can actually just "make the prototype" and the design system - instead of inspiring hundreds of useless meetings to deliberate over a flat perversion.
I like paying for things. Good things. And when I pay for good things... (and they are managed well) - they inspire people to make more good things.
This whole "Free" thing... is costing us all - big time.
For example, full-text search would be paid features, while a simple search that can be done reasonably on the client would be free.
https://www.remarkbox.com/remarkbox-is-now-pay-what-you-can....
- If you are offering a free trial, you want all your users to be customers
- If you are offering a free tier you bet your free users will either: become customers at some point or create some free marketing for you
I don't know if there are pro/cons, there are just times when you need one, and times when you need another. If you sell something really expensive, you can't put a free tier for example. It is always difficult to quantify the marketing aspect of things.
Should be noted, a limited version should still provide some value, i.e. a version of the app that doesn't save documents doesn't count because its usage in time is artificially limited by the functionality.
People seem to tend to stick and extract maximum possible value from indefinitely free plans, put up with and work around the limitations. It's a strange phenomenon but the way people evaluate $0 vs. non-$0 products seems irrational. Possibly because we can't seem to accurately estimate the time we waste on working around limitations on a free product vs. the money we would have paid for a fully functional product, i.e. estimate the time wasted in terms of money lost.
This works well in e.g. AWS where I often temporarily pay for a subcription while I'm having an issue. This weeds out for them many low-value, high-maintenance users.
(AWS in particular cannot be very blunt so it's still possible to get support without paying specifically for it... but when running a small business, please be blunt)
Also, Canva seems more like a "Publisher" killer. But Publisher was already dead.
A great question that I think must be considered. I use Photoshop all the time, but definitely don't google for it. Search trends are not a 1:1 match for usage.
That said, the Google trends for basic terms like water and food are still on the rise, even though I think they suit the question Who has to google this? even better.
Adobe has gone +50% from 2020 to 2021, and the trend seems to keep going up. I'm not attributing that to Photoshop, though.
Second, a company may have millions of users and still have no/little/less revenue. This also often would mean no support. That could be fine with a certain category. Many of the companies cited in the article are private companies, so we don't really know much about their finances.
>Microsoft Teams: Slack
>Well...they don’t. Baremetrics sold. Slack sold. You can’t compete against freemium when you’re able to offer the primary value for free. The existing market leaders can’t match your pricing without giving up millions in recurring revenue and they’re not going to do that.
Slack does have a free tier. Am I misunderstanding the paragraph as meaning that Slack lost because it didn't have one?
The freemium model can be interesting if an additional user is one entry in a database, and you don't incur that much cost from usage. For example, we make a machine learning / MLOps (oh yeah) platform at https://iko.ai. Users train models models on data, which means GPU, RAM, and storage. Unless you are venture backed and can spend money on free users to onboard a lot of users fast, it's not sustainable. We're not venture backed, and we're not really looking to be for now, which means our clients belong to a different segment.
Now, it is not excluded that the revenue generated thanks to these paying customers will subsidize free users, and we are engineering our way out of this conundrum. But someone has to pay for all that compute. One of our colleagues was supervising about thirty students doing their final year projects in machine learning, and we let them use our platform to be able to complete their projects and get their degrees. We provided support that you'd find in an enterprise offering, and we spent a lot of nights hanging out in the Slack channels helping them troubleshoot and optimize their code, giving guidance on how to train and use the models, seeing it through their graduation. You can't really generalize that level of support for all free users, and when it came to students, we couldn't just shrug it off and not support them just because they were free users, as this was personal.
As a self-funded solo founder, this has helped me gain many users since launch, which give me plenty feedback and help to validate new product features without having lots of marketing effort so far.
I don't sell any data, or display ads on free accounts. The reason to upgrade is more capacity, or complementary features for business users (e.g. alerts or organisation/team accounts).
That said, a couple of things to consider:
- Do your numbers: if you have 100 free users for every paying customer, are you going to be profitable? Can you handle the costs?
- In my experience, free users tend to attract more free users.
- They can generate a lot of support pressure, but handled right can be a great way to understand your customers problems. Handled wrong can kill you.
- Expect a small percentage of free users to cause the most trouble. From creating several accounts with different email addresses, to publicly complaining about the free plan being limited, to requesting lots of features only to switch to another free product once it's time to upgrade.
- If you're targeting B2B, price might not be as much of a factor as you might think. Are you going to save or make money for other businesses?
I could talk more about this, but will probably write a blog post one day with more learnings once I have them.
I’m happy to offer a free plan as long as it’s sustainable and I can see the benefits. So far it’s going great for me, but know what you’re getting into, and be sure you can handle the costs both in time and money with low conversion rates vs a trial model.
Edit: I moved the comment to this discussion as it's more fitting. I removed the link to my product.
You were already flagged for bad behavior there https://news.ycombinator.com/item?id=25322651
I recently shut down my free plan and instead introduced Trial on paid plans for my SAAS [1]. Brought in a new entry-level plan that is significantly cheaper to make the transition easier.
As an indiehacker, the free plan allowed me to iron out bugs and improve the overall experience of the platform. But the amount of effort it took to maintain the free plan was enormous so, I shut it down. I informed users about the decision and, I am glad few converted to being paid customers.
The most difficult part was to convert free users into paid ones. I tried everything - Coupons, discounts, emailing customers, lowering the limits, limited offers etc. Nothing really worked.
I have a lot of free time now and I can focus on delivering value to my customers. The server load has also reduced due to the reduction of free requests and database writes/reads.
For a large SAAS company like Slack, Trello etc, it makes sense to have a free plan since you have a lot of resources at your disposal but, if you are a small team or worse alone, then hell no.
[1] http://blanq.io/