Jul 19, 2024 · 49m · news
Cameron Adams: How Canva Builds Products: Lessons Learned, What Works? What Flopped? | E1179 · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this deep-dive interview, Canva Co-Founder and CPO Cameron Adams discusses the platform's core product philosophies, strategic integration of generative AI, organizational scaling tactics, and lessons learned from past failures. He explores how prioritizing user delight over standard lean methodologies fueled Canva's massive growth and shares insights on the future of AI-driven creative design.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 22% of the talking time here. How this is scored →
Harry as informed peer 4.1 Guest teaching 4.3 Guest disagreement 2.4 Harry pushing back 3.8
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 12:53 Cam pushes back on content discovery vs quality
When Harry directly challenges Cam's premise that quality content rises to the top, Cam accepts the debate invitation ('I can argue with you') and uses Taylor Swift's marketing strategy to rebut Harry's claim.
Hardest push from Harry ▶ 12:53 Harry rejects 'great content rises to the top'Harry explicitly refuses Cam's optimistic framing on content discovery, asserting that discovery saturation makes podcasting a difficult industry and directly telling the guest he disagrees.
Biggest teaching moment ▶ 43:06 Cam explains the Nature Crisis versus Climate CrisisCam clearly educates the host on the distinct difference between the climate crisis (emissions and carbon) and the broader nature crisis (biosphere, water supplies, and biodiversity collapse).
Harry holds his own ▶ 44:07 Harry counters with Oxford debate arguments on global emissionsHarry demonstrates strong macro expertise by citing Konstantin Kisin's Oxford speech, detailing statistics on UK carbon contributions (1%) and the economic growth needs of China and India.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Joining Canva: How Cameron Met Mel and Cliff | 2 | 3 | 1 | 1 | Harry opens with an anecdote from Canva co-founder Cliff regarding how Cam first met Mel and Cliff. Cam explains his background at Google working on Google Wave and how he was initially mistaken for a PHP developer candidate. | |
| Success: Timing, Luck, and the Google Wave Lesson | 4 | 4 | 2 | 3 | Harry cites Lean Startup principles and quotes early team members to ask about speed versus quality. Cam reframes luck as positioning and explains why Canva ignored investor advice to launch too quickly. | |
| Creating User Delight and the 'Aha' Moment | 2 | 3 | 0 | 2 | Harry asks how Canva creates fanatical user devotion and whether speed to delight must happen in the first session. Cam shares product onboarding details like hidden Easter eggs and triggering the 'aha' moment. | |
| Simplicity vs. Depth in Product Design | 4 | 4 | 1 | 3 | Harry quotes product leader Andrew on Cam's ability to make products 'just simple enough' and asks about the difficulty of moving into Enterprise. Cam explains the 80/20 product rule and admits their early Enterprise launch was just a sales form. | |
| Unbundling, AI Tools, and the Challenge of Discovery | 6 | 5 | 7 | 8 | Harry directly rejects Cam's premise that great content naturally rises to the top, arguing that discovery saturation makes podcasting and content tough. Cam accepts the debate challenge and counters using the music industry and Taylor Swift's marketing machine. | |
| Pivotal Product Milestones: Text to Image | 6 | 5 | 4 | 4 | Harry quotes executive guests from Spotify and Loom regarding how external AI updates disrupt product roadmaps and prompt engineering in hiring. Cam politely disagrees with the notion that prompting skills should define product interviews in the long run. | |
| AI Cost Structure, Subscription Margins, and Cloud Analogy | 6 | 5 | 3 | 6 | Harry probes unit economics, asking if heavy AI generation users erode subscription gross margins. Cam explains that bundling AI value drives long-term retention and notes that underlying AI inference costs are dropping rapidly. | |
| How AI Changes Product Team Structures and R&D | 3 | 4 | 1 | 3 | Harry asks how AI changes product team operations and what technical uncertainty Cam finds uncomfortable. Cam describes the shift toward R&D experimentation and dynamic iterations between product and engineering. | |
| Feature Rollout Strategies and Geographic Testing | 3 | 3 | 2 | 3 | Harry asks about feature rollout strategies and whether Canva uses geographic isolation like Australia or Sweden for testing. Cam outlines Canva's standard percentage cohort rollouts and targeted geographic testing for payment methods. | |
| Managing Product Change and User Backlash (The 'Glow Up') | 4 | 4 | 2 | 4 | Harry asks how Canva manages major interface changes without alienating users and presses Cam on user skip rates during redesigns. Cam reveals a surprisingly low skip rate under 1% for their major 'Glow Up' overhaul. | |
| Canva's Biggest Product Mistake: Trying to Be Social | 3 | 4 | 1 | 3 | Harry asks for Canva's biggest product failure. Cam recounts dedicating an engineer for nine months to build social network features due to investor pressure, which flopped because it clashed with private user workflows. | |
| How Canva Structures Product Reviews and AI Teams | 3 | 4 | 1 | 2 | Harry inquires about product review cadence and the size of Canva's AI talent pool. Cam shares that Canva employs around 100 machine learning engineers, largely bolstered by an acquisition in Vienna. | |
| Resource Allocation: Infrastructure vs. New Features and R&D Spend | 4 | 4 | 2 | 5 | Harry pushes for exact R&D spending percentages versus core infrastructure maintenance. Cam explains that over 50% of Canva's 4,500 total workforce is directly engaged in building and designing product. | |
| Canva vs. Adobe: Lessons on Competition and Vision | 4 | 5 | 4 | 4 | Harry asks about competing against Adobe and whether companies need a 'common enemy' to stay motivated. Cam rejects the enemy framing, emphasizing category creation for the 99% of non-designers. | |
| Quick Fire Round: CPO Swap, the Canva Rap, and BS Advice | 4 | 5 | 3 | 2 | In the quick fire round, Cam discusses the viral Canva Create rap video and criticizes generic 'Lean Startup' advice as flawed when applied without long-term founder conviction. | |
| The Climate and Nature Crisis | 8 | 6 | 5 | 8 | Harry brings up Oxford debate arguments by Konstantin Kisin on developing nations' emissions and economic growth trade-offs. Cam argues that developed nations must lead supply chain changes and help emerging economies build sustainably. |
Statements from this episode (32)
Cameron Adams: Word-of-mouth sharing has been Canva's biggest growth driver
“Launching something at Canva that people will spread, that they will tell others about has been the biggest growth driver for us.”
Adams: Canva co-founder's previous startup Fluent collapsed after $5M round fell through
“Me and my co-founders in my other startup called Fluent. We're like, very happy we were going to take over the world, and we were just about to land five million dollars in funding. Little did we know that we weren't just about to land five million dollars in …”
Adams: Google Wave failed because 2007 browsers and users were unready
“Google wave itself, like the project I work on at Google was really prescient in what it wanted to do in terms of real time communications and collaborations. But it was early for a bunch of reasons, like browsers weren't quite ready. People weren't quite read…”
Adams: Canva ignored early investor pressure to follow Lean Startup playbook
“The Lean Startup book actually launched in the year that we launched, it was like, 2012, and investors were quoting it at us, and like, telling us to launch as quickly as possible. We kind of ignored them because we knew what we wanted in the product, and we k…”
Adams: Product delight must happen within the first user journey
“Definitely needs to be in the first user journey. I don't think it has to happen in the 1:30 seconds, but I think you need to give them a feel for the product right from the very get go.”
Adams: Signup pages set the tone for the entire product experience
“If you have a crappy signup page that looks like it was made on MySpace in 2006, like that's setting the tone for your product. So they don't have to be immediately wowed, but you need to start giving them kind of the vibe of the product right from the very ge…”
Adams: Canva applies an 80/20 design rule balancing simplicity and depth
“So there's this kind of eighty-twenty rule that we apply where when you first get into Canva, you should be able to get 80% of the value, but then as you spend more time with it, you unlock the remaining 20%, which makes you an expert”
Adams: Canva's initial enterprise offering was just a sales contact page
“And what we launched three years ago wasn't really an Enterprise product. It was just a way for someone to get in contact with a salesperson.”
Adams: Single-feature AI microservices are failing due to low underlying value
“I think we're seeing, particularly in the AI space, we're seeing a lot of hype and churn and real microservices, and you're also seeing a lot of those startups kind of fall by the wayside as well as the real, you know, value of them comes to light, and it's no…”
Adams: AI content expansion won't decrease content value as quality wins
“I don't think the value decreases. There's definitely more people creating stuff, but as you very well know, great content rises to the top.”
Stebbings: Podcasting is a terrible business for newcomers today
“There are thousands of podcasts that are amazing. But podcast is a shit business to be in now, because it's so hard, unless you started years ago.”
Adams: Canva was first to scale generative images to 100M+ users
“We were first product to like truly scale generative images to over a hundred million people.”
Adams: Canva launched text-to-image AI in six weeks
“It actually only took us six weeks to launch from woe to go, from someone saying, let's do, crazy idea, let's do, like, type in a box and you get an image out of it. Six weeks later, we deployed all the infrastructure, we'd integrated into the product, and we'…”
Adams: Canva's initial text-to-image feature used Stable Diffusion
“For that one, that was based on stable diffusion and we've since, like, iterated upon that a whole ton.”
Cameron Adams: AI innovation speed will slow down within five years
“In five years time, I would say that AI innovation will have slowed by then, and a lot more will be standardized. Social patterns will be established of like how we interact with AI. The possibilities of what LLM can do are very well understood and known, and …”
Adams: Prompt-only AI interfaces in five years would show product failure
“But if in five years time, the way that we interact with AI is purely just through prompts, I think it will have displayed a total lack of imagination from us as product builders.”
Adams: LLM market will mirror cloud infrastructure competition
“There's still a huge amount of competition there, and I think exactly the same will happen with LLMs. They'll be innovating, you know, Google will add a feature, OpenAI will add a feature, some other startup that we don't even know about yet will add a feature…”
Adams: Canva bundles AI tools directly into core subscriptions
“It's all wrapped up in our Canva Pro, or our Canva Enterprise subscription.”
Adams: Paid AI add-ons will commoditize into standard tiers
“I think in the short term, it might allow them to make more money. So, like, I think Microsoft and Google both have add-on packages that you pay for to get access to their AI stuff. I think that's a very short-term thing where it'll just become commoditized, p…”
Canva rolls out new features through incremental percentage-based user cohorts
“We always roll out features as a percent, so if something is going out so we launched the Glow Up recently which is a total re, revitalization of Canva's interface. We start by rolling that out to one percent, then five percent, 10%, 20%? Hundred percent. That…”
Adams: Signposting upcoming product changes via email or social is useless noise
“More generally we try not to signposts ahead of time because it's just a lot of noise. Like you get so much email, so much social spam, a message telling you about a change that's coming is just going to fly over your head.”
Adams: Less than 1% of Canva users skip major UI updates
“Oh, not 10%. I would say less than one percent.”
Canva assigned one of its six early engineers to social features
“We had a very small engineering team. We had like six engineers. We dedicated one whole engineer for nine months to build this whole social experience. And it just didn't do very well.”
Adams: Canva employs around 100 machine learning engineers
“We've got about a hundred machine learning engineers and then, you know, product, product, product managers and designers get involved at all stages.”
Adams: Canva has integrated machine learning for over six years
“So, AI has really spiked in the last two years, but machine learning has been part of our product development process for six or seven years now.”
Adams: Product and tech development comprises over 50% of Canva's budget
“If you think of R&D as all product and technology development, be well over 50%.”
Adams: Over 2,000 of Canva's 4,500 employees focus on building product
“We've got 4500 people at Canva now. Roughly about, I think, 1700 of those are engineers, plus they're now product and design orgs with product managers, product designers. So well over, you know, 2000 folks at Canva are focusing on just building the product.”
Adams: Canva operates in a different category than Adobe
“We really consider ourselves to be in a totally different category. Like, literally. When we started, design was something that one percent of the world could do, and we wanted to bring it to the other 99%.”
Adams: Unseating a common enemy is an uninspiring company focus
“Not as your primary focus. Like if you're, if your goal is all around you know, a cervically trying to bring someone down, that's not particularly inspiring. We always focus on what is our vision of the future?”
Adams: Iteration without a strong initial vision cannot build a lasting company
“That notion that you can kind of iterate towards a product or even a company or a business model, if you don't have a strong vision to start out with, your heart won't be in it.”