Back

Office Hours: I helped scale Smallpdf to over 50 million monthly active users in 3 years and am currently the Head of Product at Jua – end-to-end Machine Learning weather forecasts. I’m Leah Tharin. AMA!Featured

Hi Elphas!

I’m Leah Tharin, Head of Product at Jua, the world’s first global deep-learning weather forecasting model.

Before Jua, I was leading the core product teams of Smallpdf a hypergrowth B2C startup out of Zurich that ended up as one of the biggest websites and fastest-growing startups in the world in 2022 during covid. I also founded 4 startups over the last 2 decades and finally sold the last one earlier this year.

I’m a leading voice around product-led growth (plg) and advise companies and senior leaders on how to get growth right when they scale. I do this over my blog www.leahtharin.com, my podcast “ProducTea with Leah”, or directly in companies as an advisor / consultant or international speaker on stage.

During my down time, I enjoy flying virtual planes in simulators, reading and writing a lot about tech, playing the guitar and cooking simple but strong-flavored things. I am a social introvert and enjoy meaningful 1 on 1 conversations and tend to get drained quickly around big groups of people.

Ask me anything about product-led growth in B2B, machine learning, senior leadership, scaling up, growth, next-generation weather forecasts, or anything else!

Thanks so much for joining us @LeahTharin!Elphas – please ask @LeahTharin your questions before Friday, September 1st. @LeahTharin may not have time to answer every questions, so emoji upvote your favorites 🔥👍🏾➕
Hi Leah! I'm a PM with a background managing infra, Dev, and ML products. I'm bringing ML to my organization and am excited to start prototyping. I've got a plan that accounts for goals, anti goals, risks, etc and have key stakeholders on board.Any tips on how to further reduce their concerns? 1:1 coffee chats to understand and reassure? Organizations without an ML practice seems to have some concern introducing a new technology.Many thanks,Cassandra
How confident are you that you understand from a scale of 1 - 10 what's important to your manager and their manager? I'm not talking about Revenue, that's obviously always important. But do you know what they are being measured on?Buy in happens through people and the positions they are in, not by necessarily addressing company goals overall. While it makes sense to have all your stuff together you could consider addressing their concerns from their level of flight.For instance, if you came to me and dropped a huge plan with antigoals risks etc. I would like that only if this is what I expected to happen. So for instance lets say we had a conversation and I said "Cassandra, take care of this ML problem and tell me how to leverage it" then I would expect that you come back a couple days later with a rough outline on how you would approach this.This doesn't necessarily have a formulated plan but outlines the structure of it:"Hey leah, i was thinking of outlining this with goals, antigoals and risks in alignment to what's important to you"From there my first question would be to give you or ask you whether you know what's important to me because it prompts me to potentially give much more business context that you might have. Which is a good thing.As an exec we oftentimes give projects and delegate things but then are not challenged afterwards on how things should be done and whether what you do is in the expectation of what I really expect. Oftentimes I just don't know myself until you ask.Could it be a 1 pager?Do I expect an entire business case?Make sure you know that before you jump into anything. The fact that you have to ask me on how to reduce their concerns without mentioning the actual concerns tells me that you might not have approached this with the concerns and already jumped into a solution.Try to make sure this is not the case and take a step back with your solution and spend some time on understanding what their fears are for the business, what they have already tried or what they are afraid of *before* you come up with a solution.It's never about AI, it's almost always about a competitor who does something in that space or what an investor said. Become an investigator before you solve the case.
Hi Leah, Thanks for taking the time to AMA! I'm the CEO of an early-stage startup (https://hardly-work.com/) and we have yet to take our big turn in user growth. We have figured out a few UX issues that we are addressing now/soon, but the sign-up urgency hasn't been there yet. For background, we are a free career development website that is bringing high-quality career development tools and services to more people. Questions: Was there a moment when Smallpdf really took off or was it a slow crawl upwards? Is consistency or experimentation better at early stages? What helped you get to your first 5K or 10K users?
Unfortunately, I joined Smallpdf already when it had 2 million monthly active users so not quite from the ground yet. However I've been part with some other 0 to 1 startups.In general thought it's less about when you reach 5k or 10k, its really more about whether you are absolutely 100% confident that you have a scaleable motion that is efficient in what you put in.So it matters more whether you got the first 5k with 100k impressions or 1 million impressions if that makes sense. Why? Because that tells you a lot about whether you should scale it already.If your motion is inefficient you should under no circumstance throw more problem at the acquisition side through paid or other means for instance but focus still on the actual product / monetization loop itself.I'm elaborating on this here in my guide about product market fit: https://www.leahtharin.com/p/leahs-product-market-fit-guideAs a general rule of thumb I think with a great product alone you can reach almost 5mio in ARR with PLG without having to think too much about Sales or crazy expensive distribution channels. After that you will almost always have to expand to higher value customers by going towards B2B or opening up your funnel far and wide.Does that make sense? I'm mentioning it here because founders have a tendency to look at users when the ratio between impressions -> Users -> paid is more important than the absolute number of the last 2 individually.
I was looking through your benchmark doc (https://www.jua.ai/benchmarking) and these results are great. An extra 20 to 30 hours of warning for extreme weather events can and will save millions of lives. I'm curious if you'd ever provide an open source version for public disaster readiness?Would you be willing to share how Jua has staffed/hired their ML Ops team, and what that process is like? Any challenges you are facing?Thanks!
That's a great question, we recently pivoted to serve the power trading business directly as this has the biggest direct economic impact globally. This means we try to provide a product that helps leverage the energy trading market directly rather than implicit services that have to translate it into value themselves.There is an incredible amount of value lost and hurting people directly every day we predict the weather wrong. Windfarms that produce when batteries are full (leading to negative prices), wind and solar plants not producing enough when we thought they would leading to critical gaps in production which leads to massively inflated prices often leaving poorer countries in the dust.This might not directly answer your question but we are about 20 people right now while the model has a comparable size of ChatGPT 3.5 so we are hyper focussed on efficiency of the individual teams. Our challenge is to keep this focus and getting the model to outperform everything else on it's core value, to predict the weather better so we don't have ML ops. We have something closely related to DevOps but not really either... they are a support team that maintains the Rest API which delivers value to the customers and makes sure our massive cluster scales with the challenges we face.I hope this answers your question somewhat. To your question of open sourcing it, I can't tell you yet but in the future it might be the case, first we need to nail the value in a quality that is not marginally but uniformly better than anything else that is in the market. And we're on the way to doing that.
How can I transition to a PM without "formal" experience, but experience in SWE, DS and VC?
You don't need to hold the job of a PM to do Product things.I always preferred Junior PM applications that showed me how they were entrepreneurs. If you look at it from the perspective of coding, instead of talking about the techstacks of what you used and built over the years show what the impact was.Whether you failed or succeeded. For instance trying to run and crash with an e-commerce store while showing that you understand what happened inside of it from the metrics standpoint is 10 times more valuable than any product certification that you can get from a school.First of all certifications only tell me that you attended something, second... Once you start working with me, I can send you to courses but I can't tell you to found a business and learn from it while working for us.Product Management is one of these disciplines where you have to think about what it is about and then you realize that you've done a ton of this already and oftentimes it's just about how you phrase and show that.Having said that, the market is quite competitive at the moment so it will also come down to a lot of persistence. And if you have never founded a little side business and looked into how a pm would run this, then you should fix that today. It doesn't need to be complicated, sitting down and thinking how we have impact and charge people for our solutions is exactly what a PM does all day.
Hello Leah!!! What an impressive trajectory. How would you say your role in Smallpdf and your role at Jua are similar and how are they different?
They are wildly different. At Smallpdf the journey was hyperscaling (which i was already familiar with) from an existing company with 20 people to 150 in 2 years while everything burns at the same time.At Jua we are much earlier and it's also ok to break things to figure out where we belong in the market and go with the company in the next couple of years. I enjoy both stages of companies but Smallpdf was very B2C / Lower B2B market focussed where as Jua is in some of the biggest Sales accounts I've ever seen.Both is exciting!
Hi Leah! Thanks for making the time :) As a fellow introvert here, I'm curious if you have any tips for how to navigate public speaking engagements. What does preparing for these events look like to you?
This is a difficult one. I love public speaking personally, so everyone is different but there are some tactical things everyone should do, no matter how experienced you are.Time your talks. Get some control over it by practicing the entire speech in front of the mirror from start to finish and measure the time.It's very rare that we go too short, we underestimate how much time it takes to get through your slides if you have them. You don't want to skip through your stuff because you ran out of time.Another thing is to direct attention. Don't show things in slides that you don't talk about. You want to be sure that people pay attention to what matters to you which is what you are talking about. People have this weird habit of cramming a ton of information into their slides which forces people to read about something you're not talking about leaving people confused.Have 50 slides with 1 sentence on them rather than 5 slides with 10 sentences on it.I do occasionally on the day still get nervous but because I practiced my material a lot before it gives me the security that I can worry about speaking and not the material so to speak. Also don't be afraid of practicing in front of a friend.Never forget, you deserve where you are and people came there to listen to you. Not a perfect version of you. Sometimes things go wrong and they do (technical issues, or you stumble on your words). If you do, pause. Take a deep breath and laugh about it.People remember memorable 3-4 quotes from talks, not your mistakes and inconsistencies. Nail those.Here's an example of a recent one I did. I was quite nervous at this talk for some reason. I didn't sleep well but it turned out much better than I thought.https://www.youtube.com/watch?v=xf6-ueNIaNI
These are fantastic tips! Especially love your last point "People remember memorable 3-4 quotes from talks, not your mistakes and inconsistencies. Nail those."I'd also love to hear specifically what you love about public speaking! Hearing these perspective always helps reframe my mindset around it.
Thanks @LeahTharin for taking the time!I’m an entrepreneur with 12 years experience in human-centered design, business development and structured finance. I am stepping down from my co-founder role next month, and am ready to go back to getting a job. In my last 2 ventures, I’ve led product and growth efforts and I’d love to position myself towards that direction, but find it tough to stand out against candidates with direct product/growth experience in a well-known company. Having been a founder of several businesses, how do you position yourself from entrepreneur to employee?
https://www.leahtharin.com/p/stand-out-in-growthproduct-with-yourA Small addition which is hypertactical. Just want to make sure you don't make this mistake. It also applies to entrepreneurs
Your superpower as an entrepreneur is not that you failed or succeeded. It is "how" you did it. It's important that you position yourself through the difficulties that you went through and how you approached the solution to them.For instance when I had to deal with Distributors for the first time in my life... I didn't know what to do. How do you structure commercial agreements with distributors? Do you just believe what they tell you? Do you hire a lawyer? what are "good" deals?The learnings are what makes it valuable, not whether my business did that right.The most distilled version of this is, don't described what you did. Describe how and what the outcome was, how many people did you affect. How did the world look like before you were there and how after (not your output). This is what entrepreneurs have ahead of non entrepreneur PM's, they also tend to get picked over classical PM's if the domain fit is better.I know this is not always easy to convey in a CV but I hope this makes sense.
Thanks for sharing your expertise, @LeahTharin! Can you share what the best CTOs you have worked with do for collaboration with you, the greater leadership team, and the company? What are they involved in, how much do they communicate with others, what decision power do they have, etc.? Thanks so much!
CTOs are always tricky because it depends in what relationship you stand to them.As Director or "lesser" PM you depend on them having a product and outcome driven mindset rather than feature factory enablement. This is assuming you don't have a CPO in your organization to fight for the product's place at the table.If a CTO is in a company with a CPO then their qualities shift more to be technological evangelists and making sure that whatever you build is in fact scaleable wherever the journey goes. In either case, a great CTO has a great product mind in the first place and makes sure that scalability and cost doesn't go completely out of control. The best ones make sure that the engineers in your organization are heard with their concerns. They are oftentimes some of the best voices when it comes to developing a product.
Thank you SO much for taking the time to reply, @LeahTharin! We do have a Product leader so the technology evangelist is where I can direct my CTO's attention.
This might be a silly question but I don't know much about PLG. What are some clear indicators that a PLG strategy would be best for a company? I would also imagine that you sometimes have to advise companies against a PLG strategy right? So, inversely, what would be some clear indicators that it would be best to seek another type of strategy for your business?
That's a great question, I'm going to do something very lazy and refer you to this post of mine:https://www.linkedin.com/posts/leahtharin_productledgrowth-growth-activity-7087006942350454785-_wFR?utm_source=share&utm_medium=member_desktopForgive me but it addresses exactly your question I think :)
Scaling a company often means scaling its leadership as well. What are some challenges you've encountered in scaling leadership teams, and what strategies have you used to address these?
It depends from where to where you scale.Critical failure points happen at different points. The first one seems to be around 50 when you have multiple product teams that don't stay out of each others business and rivalries or politics start.After 150 people stop knowing each other on a first name basis. This means you have to create "divisions" in your company and it becomes evident that you need more than just a company strategy. A product, sales and other separate strategies might become necessary at the very latest.Just these 2 stages require different skills because you're going from Startup to Scaleup, so finding product market fit which is extremely product heavy to how to grow and scale your business.A great book in this regard is "Team of Teams".