Emily Wang’s Post

We started Bento in 2020 to solve a problem we called “everboarding”. Basically, how do software products integrate a way for customers to discover the best ways to use that product, and to continue doing so as new features are added, or new use cases emerge. And over the last 4 years, we built a product that served millions of these end-users. We partnered with our incredible customers to run countless experiments and to rethink their customer onboarding and activation methods, and celebrated with them as they saw success like improving their activation rates by 4 or 5x. Yet three months ago, I made the difficult decision of winding down our operations and closing this chapter in our journey. The decision was difficult because we had over 18 months of runway and could see profitability around the corner. We had never churned an annual plan customer. Customers were seeing results and excited for their roadmaps ahead. So why shut down? We did not find a way to grow fast enough to give me confidence that this would one day be a very large, standalone company. And that was the shared goal of all the employees who joined us, and the investors who backed us, and my goal for putting 100% of my energy into Bento these past years. Surviving, or building a lifestyle business, wasn’t the goal of this particular venture. Shutting down while we had the runway also meant we could take care of our customers by offering them a smooth transition onto other solutions or internal programs. It meant being able to build out additional features to help them export data. We could take care of the team and help them land in each of their new roles. It meant returning capital to those who believed in us. It meant having the oxygen to do it with integrity and with the values that have guided us each step along the way. For founders who are in these shoes, or find themselves peering through this fog right now, it is one of the hardest calls to make. When it’s not obvious which path to take, it means you live on the knife’s edge of perseverance and stubbornness. And no founder is not without a healthy dose of both. I’m deeply grateful for the team and advisors who worked with me for weeks and months through this decision and its implications. There’s nothing I’m prouder of than what we built with our product, business, and culture. To say goodbye is deeply sad, not simply because we didn’t get to the outcome we all worked so hard for, but because the journey was so meaningful. It was, as one of the Bentosaurs put it: “a Core Memory”. Thank you to each Bentosaur, to each customer, and to the incredible investors and advisors who cheered us on, showed up to brainstorm and help, and opened countless doors. This journey would have been far less meaningful without you. As for me, I'm excited to be exploring at the intersection of climate & energy and finding the next great mission to embark on!

Our mission at Bento isn’t just to arm PMs and Designers with the power to create magical onboarding and activation flows – 𝐢𝐭’𝐬 𝐭𝐨 𝐡𝐞𝐥𝐩 𝐭𝐡𝐞𝐦 𝐛𝐞 𝐬𝐮𝐜𝐜𝐞𝐬𝐬𝐟𝐮𝐥, and in turn, unlock meaningful improvements in conversion metrics and user experience delight. But what does it mean to be successful? 🏗️ In the early days, it was about giving people an 𝐢𝐧𝐭𝐮𝐢𝐭𝐢𝐯𝐞 𝐛𝐮𝐢𝐥𝐝𝐞𝐫 𝐞𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞 🎨 Then it was about giving them deeply extensible 𝐜𝐮𝐬𝐭𝐨𝐦𝐢𝐳𝐚𝐭𝐢𝐨𝐧 that would also “just work” with clever defaults. 🛫 Then we made governance and scaling easy with infrastructure like 𝐀𝐢𝐫 𝐓𝐫𝐚𝐟𝐟𝐢𝐜 𝐂𝐨𝐧𝐭𝐫𝐨𝐥 But lately, our focus has turned to “𝐝𝐞𝐛𝐮𝐠𝐠𝐢𝐧𝐠”. Why? 🕵️ The more we invite non-engineers into the traditional world of engineering, 𝐭𝐡𝐞 𝐦𝐨𝐫𝐞 𝐰𝐞 𝐡𝐚𝐯𝐞 𝐭𝐨 𝐚𝐫𝐦 𝐭𝐡𝐞𝐦 𝐰𝐢𝐭𝐡 𝐭𝐡𝐞 𝐬𝐚𝐦𝐞 𝐤𝐢𝐧𝐝𝐬 𝐨𝐟 𝐛𝐮𝐦𝐩𝐞𝐫𝐬 𝐚𝐧𝐝 𝐭𝐨𝐨𝐥𝐬. So far, we’ve invested in tools that help customers answer questions like: 1️⃣ Why isn't a guide showing up (are targeting rules set incorrectly? Is the guide even launched?) 2️⃣ Why isn’t a guide appearing where you expect it to (are wildcards used correctly? Are you testing on the right URL? Is the target element even on the page?) 3️⃣ Is an element you’re tracking for auto-completion a reliable one? (and can we suggest a better one?) These investments are critical to giving our customers confidence in relying on Bento to power some of the most valuable parts of their user journey. And they're already saving us a ton of time supporting customers when they have questions. Stay tuned for more to come!

🎙 Excited to share my podcast episode on Growthmates with Kate Syuma! If you don't know Kate, she was the former head of growth design at Miro and is an incredibly thoughtful and creative thinker. 🎧 Our conversation spanned a lot of topics, but a few things I hope you'll enjoy: 1️⃣ Embracing the power of culling features: We discussed how letting go of certain features isn't a failure but an investment in the greater product. It takes discipline and a clear idea of the longer term product vision. And as such, it should be celebrated! 2️⃣ Finding intrinsic motivations: This is true in all roles, but particularly as a founder, it's crucial to anchor to a purpose and goal that isn't as easily swayed by all the ups and downs of external validation. 3️⃣ Many strategies are valid...if applied appropriately at the right time: Of course we dove into PLG (Product-Led Growth) and more broadly, self-serve. Knowing when to use these levers, and for which customer segments, can make all the difference between "checking the box" and really learning and iterating. 🚀 Thanks for having me Kate and excited to hear what the community thinks! #PLG #Growth #Entrepreneurship

A few weeks ago, we previewed a feature where customers can describe their customer segments using natural language and have Bento + GPT turn it into structured rules. ✨Excited to share it’s now in Beta! ✨ But before we decided it was ready, we really wanted to take a swing at reducing perceived slowness. As context, the average prompts would return structured rules after 8-15 seconds. - For some, that’s fine. But if you’re impatient like me, that’s just awful. - And unlike text responses we need to return structured rules, so that prevented us from streaming in the responses one token at a time - But the worst were large customers with tons of attributes in our system. They could experience up to 30 seconds of wait time 😱 So here are some things we’re trying. 1️⃣ Better than blank. We added in a classic “progress bar” / “spinner” treatment along with some text describing what’s happening under the hood.  It’s better than a blank screen, but pretty superficial as a “solution”. 2️⃣ Play with psychology. We realized that being stuck waiting was almost as bad as the wait itself.  Now, after ~8 seconds, we show a prompt that asks the user whether they’d like to keep waiting, or abandon and write the rules manually. The interesting thing is that it takes the mind 2-3 seconds to read, process, and decide ⏲️, so oftentimes the results would show up right as you decided to keep waiting. 3️⃣ Experiment with pre-processing. Unsurprisingly we realized the biggest contributor to latency was the number attributes we passed into GPT. We’re continuing to experiment with different ways of pre-processing and pass in fewer higher quality attributes. If you’re also working on applying AI to your product, and thinking through the hairy interaction challenges, I’d love to connect! #everboarding #growth #ai

Startup life can be super heads-down. But moments like this make us (literally) pick our heads up and crack a smile. Feels like a dream to be in Time Square (and next to Chicago no less!) this week. Vanity metric? 💯 Reminder to have fun along the way? 💯💯💯

The most exciting to us at Bento isn’t actually the new product capabilities we ship – it’s the impact on our customers’ ability to scale themselves to unblock their users ✨ Patrick and the team at Evisort beautifully integrated Bento into their contract management system to offer users high quality contextual guidance. What does that mean? ✨  It’s clean and consistent across their application. 📺  The content strikes a great balance between high level “here’s why you should care” and easy ways to dive in. 🧑💻 It’s designed for the admin-user who wants to self-serve. And the data proves it! Of the users engaging with this contextual guidance, most use it repeatedly. It’s a great reminder that #everboarding truly isn’t one-size fits all. To build a “new user experience” or “self-serve” flow that works for everyone is unlikely. That’s why having the flexibility in a platform like Bento can unlock many different approaches, each targeted to different user segments and profiles 🎯 versus being locked into a single hard-coded experience. Check out the story in the link below!

Sneak peek of an exciting new “quality of life” feature we’re working on: generating guide targeting rules automatically! Figuring out the proper construction of AND/OR rules and attributes + operators can be a little daunting for non technical builders, so we’re leveraging GPT to take some of that friction away. Before we get this into users’ hands, we’re still working through things like: 🏎️ How to make this faster. Latency is a big part of the equation for whether “AI features” are “worth it” and things are still too slow right now. 🧠 How to give users actionable next steps when GPT returns rules for data they don’t have. 📈 Measuring and instrumenting each attempt so we can thoughtfully iterate! excited to share back when it's live!

It’s always a little scary when your designer wants to do a “redesign” 😅. Yet deep down, we know a series of small fixes and bandaids 🩹often aren’t enough. I’m excited to start off 2024 with a thoughtful redesign of a core Bento page that houses our “engines” (aka Guide launching rules) ⚙️⚙️. Because Bento guides can be so flexibly triggered through rules, manual launching, or via other guides, this can be a tricky area for our customer to navigate. I’m even prouder of how Andy, our designer, jumped at the opportunity to write his first ever blog post (link in comments)! It dives into how the redesign came about (surprise, it was actually supposed to be about AI!). It’s hard stretching beyond your usual “skills” but the best teammates see it as an opportunity to learn and grow. Congrats Andy!

I'm a bit of an airplane/flight geek. My social feeds are regularly filled with videos of pilot radios, and "close calls" where two planes narrowly miss each other 🛫 🫣 In a world where so much of flight is automated, we still rely heavily on #airtrafficcontroller (ATC) and their underlying systems to make sure the thousands of people flying every day don’t end up on a collision course. Recently, I logged into a tool and got a new feature announcement modal on top of a tooltip, ~7 "new" labels, and a carousel of upsell cards 😱. (I'm pretty sure I came in just to add a new user to our account) But these things happen all the time!! The (many) teams building and launching in-app experiences lack a system for controlling them in aggregate. Which is why I'm so excited to share the release of Bento's own ✈️ Air Traffic Control system ✈️ Over the past year, the number of components and experiences we support has increased significantly (especially with Surveys and Flows). And so we majorly overhauled our system for controlling #priority #visibility and #throttling. I'll share a blog post below 👇 diving more into Bento's ATC and Command Center. And I'd love to hear how you've wrangled the chaos in your own experiences!

if you're looking for 10 min of procrastination on this lovely Friday, I'd love your feedback on our new homepage! 💭 Messaging is hard, and we're taking a new approach to make our #differentiation clear – is it landing? 🎨 Playing with some new animations to bring the essence of the product to life – does it work for you? 📋 Survey here: https://lnkd.in/gSQSyeKh   The first 10 respondents will get a Peet’s/Starbucks coffee card on us!