Reckless assumptions with a sprinkling of AI on top 💩

👉 have you ever done user needs analysis? ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ Hey folks, I just scrolled two miles on LinkedIn to find a [post that I'd liked a couple of days ago]( (and recklessly didn't save), because the search functionality is so so bad: [Image] ...but instead of looking into that, and gazillion other bugs, LinkedIn decided to remove the "curious" (🤔) emoji. How? How did they decide this? Why did they decide this? Did the users need it? Well, it may or may not surprise you, but even huge companies are about just as YOLO in making product decisions as your 20-year-old brother working on his startup in your parents' garage. Customer [needs & wants analysis]( anyone? Pfff...let's just slap some AI over it! [Thumbnail for video]( [Watch this freshly-squeezed video rant!]( Countless examples of product companies just playing their hunches are jumping out at me recently. I especially liked the Duolingo one from [Lenny's newsletter](, because it was...well, cute 🙈 I wonder how many product decisions are based on product teams' preferences? Here's how Duolingo tried to increase their stickiness and accelerate their growth: [Image] [Image] [Image] 😬 So making a completely arbitrary assumption about customer needs based on your own, personal preferences is one way to fail at a product launch. What else? 2. Misunderstanding the root cause behind your customers' motivation to act: [Image] This fragment comes from Nir Eyal's [recent newsletter.](  People don't act because there's some reward waiting for them at the end (sorry, success messages!) - they act because they want to avoid pain. How is the new feature in your product going to solve their pain? Sequel's recent AI addition is an excellent example of a feature that is both trendy and solves the users' pain of having to repurpose the webinar content into blog posts/ social media summaries: [Image] What else? Quite obviously now... 3. Blindly following the trends As Digiday pointed out[in their recent article](: "The flurry of news related to artificial intelligence might soon require people to have their own AI to just keep up with — and make sense of — all the updates." Slapping AI-based features on top of your product without doing a proper needs analysis is not a silver bullet to Product-Led Growth, unlike what many product teams now seem to be thinking. It may also actually...lower your accuracy and tarnish your reputation if you make your users believe GPT-4 can "answer" their questions reliably. As Intercom wrote about their recent release: "The trickiest thing about large language models (LLMs) is that they’re great at appearing plausible, even when they’re wrong. ChatGPT suffers from “hallucinations” – confidently providing incorrect information." How to get a better idea of what your users need?  1. Segment customers to collect data from the right target audience Before you send your users surveys to collect product ideas, you should first segment them so you can understand how the needs differ between different user groups, cohorts, or personas. With Userpilot, you can [segment customers]( in multiple ways, such as: - Feedback responses - Activity - Personal information - Company data - and Engagement profiles  [A screenshot of Userpilot showing how you can segment by different options for customer wants and needs analysis][Userpilot]('s segmentation options. This will help you examine each segment in more detail and uncover unmet customer needs. 2. Ask them. In-app [surveys]( and NPS surveys can help you assess the customer’s needs at different stages of the user journey, and give you a better idea of what may need improving. [A screenshot of Userpilot showing a in-app survey used for customer needs and wants analysis] You can easily build these in-app surveys with [Userpilot]( too (and just bear with me - they are currently getting a massive face lift 🔥 sophisticated branching logic, standalone surveys...I didn't say anything 😶)  3. *Plan* your experiments - grab [Ben Williams']( [experimentation template]( Should go without saying, but as we all know - this "always start with a WHY" thing is easier said than done...so Ben helped by kindly sharing his experimentation template, so once you know what the users say they want, you can quickly (and cheaply) check if the users really want it. [Image] He shared more [goodness on LinkedIn recently](, including an excellent Monthly Impact and Learnings review template: [Image] 4. Run [a fake door test]( followed up by feedback Fake doors testing is a technique used to assess market demand for a product before investing in its development. So it can save you major bloopers like Duolingo and months in wasted resources. 5. Analyze and visualize your experimentation data Obviously after running the experiments and surveys, you still need to aggregate and analyze the data by segment. So when picking your survey/in-app experimentation tools, make sure they come with proper data analytics: [A screenshot of Userpilot's NPS dashboard] [Userpilot's]( PMF analytics dashboard - coming soon. And that's it for today! P.S. Last chance to sign up for our [LIVE webinar on SaaS In-App Messaging]( - grab a seat and learn how you can use in-app messages for [customer wants & needs]( analysis! [Image]( [Image] Emilia Korczynska, Head of Marketing at Userpilot I'm a marketing manager obsessed with product growth. Wanna talk? Simply respond to this email! To make sure you keep getting these emails, please add emilia@userpilot.co to your address book or whitelist us. Want out of the loop? Don't remember you subscribed at all? We get it. We sometimes don't remember how we got to our office today let alone how we subscribed to this or that email. Sometimes people also get offended by our strong opinions on all matters product, SaaS and UX, but you know what? We won't stop sharing them - and what we believe is the best product practices and the future of SaaS. Anyway, if you ever want to come back you'll know where to find us. Until then! [Unsubscribe](. Our postal address: 1887 Whitney Mesa Dr #9995 Henderson, Nevada 89014 United States