Jonas O. Klink is Vice President of Product Management & Design at Thrive Market, a healthy and sustainable online grocer. He started his career as a software engineer and transitioned into product management while at Google, where he worked on assistive technology for Google Toolbar 5. Jonas later worked in product and leadership roles at eBay, Walmart, and WeightWatchers, where he focused on digital shopping and ecommerce.
In our conversation, Jonas talks about his team’s initiatives to “better the learning velocity” — taking an initial idea through hypothesis-driven development to build customer-centric, scalable solutions. He shares how the mission behind Thrive Market — to make healthy and sustainable living easy and affordable to everybody — informs their strategy and acts as a North Star for everything they do. Jonas also discusses how Thrive Market wants to become an industry thought leader and is rethinking the search and browse experience for online grocery.
Thrive Market is a healthy and sustainable online grocer. We offer a membership model that allows members to access thoroughly vetted food and household products. Our members benefit from the convenience of having access to unique brand products only available from us and some of the best national brands that all have been heavily vetted. Our merchandising team thoroughly vets products on our marketplace to ensure they don’t include anything from our list of banned ingredients and adhere to our high-quality standards. Our goal is to carry the healthiest and most sustainable products in the food and household space.
Our approach is a relatively conventional model of an ongoing grocery service. When you place your order, it appears at your doorstep a day or two later, and you can subscribe and save further on those staples you love.
We recently significantly upgraded our entry experience for new members by enhancing the quiz you take when you join. That way, we can learn more about your household, your dietary and value-based preferences, what kind of foods you like to have on hand, and your tastes, preferences, and cuisines. We then use some AI on our backend to generate a cart based on your personalized preferences, but that cart can be fully edited and customized through conventional search and browsing.
When you place your first order, it’s a fully customizable service. You can choose to have weekly, bi-weekly, monthly, etc., deliveries that will show up on your doorstep with the products you select at a further discounted price. Or you can shop ad hoc and place an order whenever you want. It’s a unique blend of food subscription services that offer complete control over what you want when you want it. There are no shopping windows, blackout times, or constraint catalog. The whole catalog is available to you any time to shop, just as if you’re shopping at any other non-subscription online grocery.
When Nick and Sasha, our co-founders, approached me with this role, I was excited because the mission is near and dear to me. I’ve spent a lot of time working for large companies and wanted to do something that aligned with my personal passions: health and fitness for myself and my family, and sustainability. This was a very natural fit from a mission perspective, and they gave me free rein to stand up a proper product and UX design organization.
It’s challenging to change practices at Google, eBay, or Walmart. I’ve done what I can to incorporate good practices and, coming into this role, I was given carte blanche to do everything the way I wanted to. I’ve been working hard to ensure that we have the right talent in the right places, that our employees are supported and upskilled, and that we’ve put together a world-class product management and UX design process.
I spend 80 percent of my time thinking about the vision and strategy. Another 80 percent — and I know the math just doesn’t add up — is making sure that I’m supporting my team in how we go about things. They’re all extremely talented, but I need to be there to help them make the right decisions and move every opportunity forward.
As a leader, I have two main jobs: to be very clear about where we’re headed and how we measure progress toward that, and to set up the people executing those goals for success. That means making sure that we have the right talent, making sure our employees have the right career paths, evolving how we improve the process of talking with our members, and making sure that we scrutinize the problems we choose to execute. Because we’re a small team, everything we invest in must be well thought out.
It’s a combination of art and science. On one hand, we are a business. We’ve built a framework as a public benefit corporation to balance purpose with profit, ensuring we meet top-down goals to satisfy specific business needs regarding growth and profitability. Those needs can help meet bottom-up goals anchored in the vision I’ve put together as a function of our mission. Part of our annual planning and constant check-ins is ensuring those two meet in a healthy place, and we use the OKRs to do so.
That’s where it gets enjoyable and interesting to me. Our primary job in the product and UX team is to advocate for our prospects and members. We make sure we understand what we can do for them, but adding in that business layer, we also have to solve these problems in a way that aligns with the business at scale.
It starts with aligning the vision, strategy, and goals to minimize those conflicts as much as possible. I’ve been part of companies where there’s part of the company that does the mission-related work, and then everybody else is making money for the shareholders. I reached a point in my career where I wanted to find a place where you could merge the two, which led me to Thrive Market. I’m very pragmatic, so I’m rooted in measurable things. If you really want to be a mission-driven company, you should treat your mission not just as some company letterhead, but as a North Star.
One of the first things I did was ask many questions about the mission itself. It looks deceivingly simple to make healthy and sustainable living easy and affordable for everybody, but how do we define healthy? Do our design targets share that definition? How do we define affordability and assign some measures toward those? Are we making progress every quarter toward realizing that mission? That informs my product vision — a two-to-three-year construct to move our product offerings closer to our mission in a way that’s aligned with the business. That translates to a strategy — typically a one-year construct that then translates to a set of focus areas or priorities.
It’s rooted in the realization that simple statistics exist around a team’s ability to distinguish good from bad ideas. When you look at the research, they quote a win rate of somewhere between 10 and 40 percent. Typically, that’s tied to how long the team has been together and how much they understand the particular problem space. So roughly one out of five ideas is going to be successful. How do you know which are the good ideas and which are the bad ideas?
For the average person, we have to play the statistics. It becomes a numbers game. You have a win rate, which we actively track, of about 30 percent. Your goal is to produce and test ideas as quickly as possible and generate clear learnings. I introduced a hypothesis-driven model. We start with the well-understood problem statement that’s been vetted in quantitative and qualitative data. Once you have a good understanding of the problem, you have a conversation about what is potentially causing this problem.
You enumerate all possible hypotheses and prioritize the ones you want to pursue first, and that feeds into an art form of testing 80 percent of hypotheses with 20 percent of effort. Once you’ve proven or disproven the hypothesis, you can invest and build on a full-blown solution. At the end of the day, it’s also a mindset shift in the sense that we’re not building features. We are an efficient machine for proving or disproving hypotheses. Once we’ve proven one, we can build the solution, simplifying discussions.
We found that at a company of Thrive’s size, and given the catalog is so heavily curated and much smaller than many of the catalogs I’ve worked on, many of the tactics and the ideas we were testing needed to be fixed. We took a bunch of smart people who have collectively all worked on ecommerce, and made enhancements that worked from our past experiences. We ended up not moving the needle barely at all. We could have made one page better, but we stole performance from other parts of the funnel — this was a great learning that there are no shortcuts in this particular space.
It’s been fascinating to see that while you have a robust playbook and experiences, there’s no substitute for genuinely understanding your design target and the challenges it faces.
First and foremost, everyone working on your team needs a clear understanding of what’s expected of them. I can’t tell you how often organizations get in trouble because they say, “Career ladders seem like fluff. How big of a difference do they make?” Ultimately, there’s a social contract between the company, the people manager, and the individual contributor. Creating career ladders is step number one.
Step number two is to look at these goals — these areas of ownership — and evaluate if they’re setting each individual up to showcase the skills I expect them to have, and this has to happen before you do goal setting, OKRs, and so on.
One of the big takeaways I’ve had from some leadership training is ensuring that feedback is given in time and is very specific. I’ve given my team permission to call me out if I ever say something, even if it’s just positive. If I say, “Great job on that presentation,” they can ask, “What was great about it?” That way, I write feedback right after it happened, tell them what was good or could have been better, and give them concrete examples. I won’t remember six months later, so why would I wait? It’s doing everyone a disservice by not giving them coaching.
I’m most excited about getting to the depth of some of the more complex parts of our experiences in our business. I consider a lot of the things I did last year very foundational. I spent a lot of time laying the groundwork for ways of working, evaluating talent, ensuring we have the right team, and plugging obvious holes in the experience. Now, we’re getting into exciting areas of pushing what online grocery can be and should be. As we focus on healthy and sustainable living, online grocery has a lot of promises left to fulfill.
By talking to members, online groceries do not solve many of the actual problems. When you go to Amazon, you build a cart of two or three things and searching and browsing work fine. But with groceries, you’re trying to shop for a family for the week. I have two kids, one in college, and two dogs. There’s a lot of us. There are way more than a few products in that cart. I’m challenging ourselves to think, “Yes, search and browse is familiar and there’s value in that, but is this the best way to build large grocery carts?”
How can we continue to make this service great for Thrive customers and be a thought leader in the online grocery space? There’s a lot of work left to do. This year, we are starting to think about how we can really challenge the status quo. We’re in a unique place to do that.
To me, generative AI, while it’s super cool and has many interesting aspects, is just a tool. The tool is only as good as the problem that it can solve. We’ve experimented with how it can help with personalization and reduce the cognitive burden for our customers of building their carts and finding the right products. The value of a large language model is that it essentially knows the whole Internet, but the problem with the large language model is that it does not understand our catalog. Our catalog sits behind a paywall.
It knows the Internet well and knows people’s general tastes and preferences. If you tell ChatGPT, “I’m a family of two kids and two dogs. These are our preferences. Can you build me a meal planner for the week?” It can do that well. But it doesn’t know our strengths. It would end up recommending products that may not be available or aren’t a great match with what our members enjoy.
The next step is to have a localized version that can access our catalog and better understand those pieces. So far, we’ve found that more traditional machine learning techniques built by our team are delivering a lot of value to our customers.
LogRocket identifies friction points in the user experience so you can make informed decisions about product and design changes that must happen to hit your goals.
With LogRocket, you can understand the scope of the issues affecting your product and prioritize the changes that need to be made. LogRocket simplifies workflows by allowing Engineering, Product, UX, and Design teams to work from the same data as you, eliminating any confusion about what needs to be done.
Get your teams on the same page — try LogRocket today.
PMs don’t need fake authority to lead well. Learn how shared context, clearer trade-offs, and better decision-making build stronger teams.
Learn how PMs can spot novelty effects in A/B tests, validate wins over time, and avoid mistaking short-term lifts for impact.
Map AI data risks, vet vendors, run safer pilots, and build legal buy-in for AI tools without creating security gaps.
SVP of Product Sriram Iyer visits to chat about how he uses AI to launch the “thinnest slice of pizza” product and shift mindsets around AI.
Learn more →
