Vitaly Friedman’s Post

What Does “PM” Actually Mean? The difference between product managers, product owners and program managers? ↓ For a long time, I was utterly confused about the term “PM”, and the roles and responsibilities that go along with it. I would jump from one company to another with wrong assumptions and expectations — until I’ve noticed that companies often define the role of a “PM” very differently. So I would gently navigate around and try to subtly figure out their exact tasks. Until I stumbled upon what is probably the clearest definition I’ve seen so far in “Build”, a wonderful book by Tony Fadell: → PRODUCT Managers are the voice of the customer. Their focus and responsibility is to figure out where customers are unhappy, what product should do, create the description of how it will work, and the facts they want customers to understand. To achieve it, they work with every part of the business — from engineering to finance. → PROJECT Managers are the voice of the project. They coordinate tasks, meetings and assets to enable individual projects to get done in time. → PROGRAM Managers supervise groups of projects and project managers, facilitate process and make sure all projects across teams get done in time. → PRODUCT Owners work mostly with the development team to prioritize the product backlog and create user stories. In some way, they dictate what will and what will not make it into the product. (Which is something that actually should be done together *with* the team, rather than *for* them.) Most teams assume that PMs are product managers, and often POs and PMs are seen as the same role — and sometimes they are, but often they aren’t. So when I hear the abbreviation appearing in a conversation, I instinctively ask specifically what we mean there. Program managers are not the managers of product managers. Project managers aren’t managed by program managers, but they do report their progress, as program managers coordinate projects across teams. And most importantly: don’t assume a specific set of skills and responsibilities from a PM role alone — they are often shaped and adjusted by a PM, their expertise, experience, skills and interests. So ask and double check to avoid misunderstanding down the line. #ux #design

🐉 Designing Accessible Drag-and-Drop UX (https://lnkd.in/dHEqHtT5), a practical case study on how to design and build drag’n’drop for cards, table columns, resizing and reshuffling with keyboard support — on mobile and on desktop. By Alex Reardon and Atlassian team. Open-source library: https://lnkd.in/dsGhEjhK Accessibility: https://lnkd.in/dKqjBpJ8 Docs: https://lnkd.in/dKJFEhwh Examples: https://lnkd.in/dK2tu2dz ✅ Users should feel like they move physical objects around. ✅ Design “grabbed”, “dragged”, “dropped”, “error” states. ✅ Design the “activated” and “disabled” drop zone areas. ✅ With cards, remove the item from the stack on drag. ✅ With tables/grids, remove the item from the stack on drop. ✅ Dragged items should move towards users in z-dimension. ✅ Dragged items move content out of the way to make room. ✅ Animate the drop of an item into its new position (100ms). ✅ Add shadows and elevation to make interaction obvious. ✅ Simulate a “magnetic” effect that snaps objects into place. ✅ Reshuffle when a center of a dragged item overlaps an edge. ✅ Keep handle icons accessible and test with screen readers. ✅ Spacebar to pick up, Arrow keys to move, Space to drop. ✅ Use grab-handle icons and adjust system cursors. ✅ Use a haptic “bump” to indicate grabbing on mobile. In light of European Accessibility Act and WCAG 2.2 AA it implies, many companies struggle to implement accessible drag-n-drop in their products. In fact, most solutions out there rely on mouse interactions alone — without keyboard support, and hence aren’t accessible. Once we have reshuffling columns in a data grid, resizing, map interactions or moving cards on a board, we need to design handles and UI controls that support navigation with arrow keys, Esc, Enter, Space. Atlassian’s open-source library is a good option to start with, but intricate drag-n-drop interactions will require a bit more work. Useful resources: Atlassian Design System (Figma kit) https://lnkd.in/dqtiVFQs Atlassian Core Components + Sticker Sheets (Figma kit) https://lnkd.in/dXuN67Bi Drag and Drop UX for Design Systems, by Grace N. https://lnkd.in/e-Q-qBPe Drag and Drop UX Best Practices, by Ceara Crawshaw https://lnkd.in/dNdP-iyr Accessible Drag-n-Drop, by Jesse Hausler https://lnkd.in/es4HhyR9 Goldman Sachs Drag-n-Drop https://lnkd.in/de9vZN7q Practical Guide To Drag-n-Drop UX, by yours truly https://lnkd.in/dX-6jU9N #ux #design

✨ What Designers Need To Know About AI and LLMs (https://lnkd.in/ehSSRp2j), a very practical and easy-to-follow overview of all the inner workings of LLMs, Artificial Intelligence, machine learning and deep learning — from zero to ChatGPT. Kindly put together by Andreas Stöffelbauer. ✅ Large Language Models (LLMs) are built on machine learning. ✅ They have one goal: predict the next word in a sequence. ✅ LLMs assign probability to words, then choose the “best”. ✅ Temperature defines how much LLM can deviate from it. ✅ High temperature (0.8) → creative, low (0.2) → precise. 🤔 LLMs are trained to generate human-like text, not true text. ✅ Nothing in the training signals truth or reliability of data. ✅ Content on the web sounds confident, so LLM does too. ✅ Way out: extract context from web search, pass it to LLM. ✅ Training takes time (3–12 months), so LLM has a cut-off date. 🚫 LLMs can’t reason, and they don’t build a knowledge base. ✅ Every LLM has a system prompt with limitations as fail safe. ✅ Context windows define how much LLMs keep in memory. ✅ As users continue a conversation, LLMs start forgetting things. 🚫 You can’t generate good output without good training data. While some nuanced questions might be immediately clear to humans, AI often struggles to notice small details that make a huge difference. In fact, by default, LLMs are not aligned with human intentions, so training and fine-tuning are very much needed to produce output that’s closer to what humans want and expect. In their essence, LLMs merely model natural language with a large number of parameters. They generate the next word, one word at a time. But to do that well, they need a *vast* amount of clean data to be trained on. And that training is remarkably expensive for the environment. Training GPT-3 alone produces as many CO2 emissions as full 28 years of a person’s life. So before adding an AI-feature to your product, think twice. No AI sparkle can magically overcome years of design and data debt. You will probably spend quite a bit of time with information architects and data scientists to gather, prepare, organize, clean and delete data. LLMs average remarkably well, but they rarely pay enough attention to outliers until you point them in that direction. They work best when they are heavily scoped to a specific context, documents and sources — so we can use oour good old-fashioned search to feed AI with context on what they need to summarize, visualize or explain. As Josh Clark noted, AI engines are mostly dream machines. And so they respond poorly to critical tasks that require a high level of accuracy. Don’t treat AI output as oracle, but rather as a temporary artefact with knobs and presets to allow humans to refine it fast. It might be good enough, or not, but its value will come from user’s critical thinking and a broader understanding of reality. #ux #design #ai

🔬 UX Research Framework (PDF, https://lnkd.in/eRRwd4nW), a visual guide on how to apply UX research for intentional design decisions, even when under pressure to deliver. Start at the top left, and move through the diagram to organize your research. By Erika Hall 👏🏼👏🏽👏🏾. 🚫 Research questions are not interview questions. 🚫 You can’t just ask people what you want to know. ✅ The key to good research is finding good questions. ✅ You won’t find them until you admit you have no answers. ✅ Establish clarity first: what you need to know, and by when. ✅ Before brainstorming ideas, brainstorm useful questions to ask. ✅ Keep a running list of relevant questions for entire company. ✅ Turn highest priority questions into UX research projects. ✅ How you phrase questions shapes what you can learn. ✅ Set goals: real-world check, find bottlenecks, explore ideas. ✅ Explain what data you need to be confident in your decisions. ✅ Explain how you will gather data, timelines, roles and costs. 🚫 Spot and flag decisions that are based on guesswork. ✅ Research methods are ways to answer your questions. ✅ Choose methods that will give you the data that you need. ✅ Good research is always rooted in the past, not the future. ✅ Keep adding emerging questions to the list of known unknowns. As Erika Hall notes, too often companies invest an incredible amount of effort in research, just to leave it in a research repository on SharePoint, never to be touched or used. Look ahead how you will document, present and integrate your insights. Set timelines, roles, metrics and a reporting schedule. Companies typically block research not because there isn’t enough time or money, but because they think it’s expensive, time-consuming and will cause disruptions. Asking questions is inherently threatening to authority, and management often fears being undermined or questioned. There is always time and money for research. You can always learn something useful within whatever time and budget is available. Address high-priority questions first. Help teams integrate insights. Establish relationships. Flag assumptions. Track the impact of your work. Start with a small commitment, then grow: slowly but steadily. Useful resources: How to Choose The Right UX Research Methods, by Allison Grayce Marshall https://lnkd.in/eyKW8nSu When to Use Which UX Research Methods (+ PDF), by Christian Rohrer https://lnkd.in/eJY7yCHW Research Questions Are Not Interview Questions, by Erika Hall https://lnkd.in/ef_8jNgU How To Build UX Research From Scratch (+ Templates), by Ori Dar https://lnkd.in/eV533Qkt 🪴 UX Research Field Guide (Free Books) https://lnkd.in/eW8xGebQ #ux #research

🧭 Design Principles (https://principles.design), a wonderful collection with 195 pointers for design principles and methods, searchable and tagged, from hardware and infrastructure to language and organizations. Kindly put together and maintained by Ben Brignell. 👏🏼👏🏽👏🏾 Useful examples: Guiding Principles for Experience Design, by Whitney Hess, PCC https://lnkd.in/edavNGiF Ten Principles for Good Design, by Dieter Rams https://lnkd.in/e_hrSkH4 Principles of Product Design, by Joshua Porter https://lnkd.in/eQjWf6nT Design in the Era of the Algorithm, by Josh Clark https://lnkd.in/egfKmez8 Designing for Voice Interfaces, by Brian Colcord https://lnkd.in/e5Gca3Ta Gov.uk Design Principles https://lnkd.in/efd8UZdX Chatbot Design Principles, by Emmet Connolly https://lnkd.in/e4xD4zes Voice UX Principles, by Ben Sauer https://lnkd.in/dQACgwue How To Establish Design Principles ↳ https://lnkd.in/eTfDtQQB, by Marcin Treder ↳ https://lnkd.in/e9JwbaKb, by Better 🧲 Design Principles in Design Systems: Audi: https://lnkd.in/eDeC2VsW Atlassian: https://lnkd.in/efFjUw7X 👍 Carbon: https://lnkd.in/eAqGd4yc Intuit: https://lnkd.in/ezHG5bPp NHS: https://lnkd.in/ev-t_QPB Nordhealth: https://lnkd.in/e9NdzVsX 👍 Spotify: https://lnkd.in/en-_-dHa USWDS: https://lnkd.in/emjCzmr9 👍 Zendesk: https://lnkd.in/eZjNsqep 👍 Design principles are often seen as very rigid and uncompromising guidelines that dictate design decisions and specific ways of working. But actually, they are an incredible tool to rally the team around one purpose, but also document values and beliefs that the organization embodies. It aligns teams and informs decision making. Most importantly, it helps avoid never-ending discussions that are often a matter of preference or taste. But design shouldn’t be a matter of taste. Good design should solve real challenges. And we should be able to assess how well a particular design solves a particular problem. Design principles help us get there — and avoid debates that might not matter that much in the first place. Do you have design principles established by your design team or your organization? Please leave them in the comments! 🙏🏽 #ux #design #designsystems

🎙️ For the weekend: Future Of UX and Digital Design (https://lnkd.in/epd3JaTP), a friendly video/audio podcast episode with me, hosted by wonderful Chris Do — on the present and future of design, useful UX strategies and insights for more inclusive, exciting and respectful digital experiences. Podcast (Apple, Spotify): https://lnkd.in/e_9hCjg7 Video (YouTube): https://lnkd.in/eAYrtRKa 🎙️ Bonus: Useful Podcasts For Designers https://lnkd.in/eCWZT5hq --- A *huge* thank-you to Chris for welcoming me to the show, and I can only wholeheartedly recommend Futur as a podcast on everything design, business and career. Perhaps a wonderful listening material for weekend errands! Happy listening (or watching!), everyone! 👏🏼👏🏽👏🏾 #ux #podcasts

⚡ How To Ask Users For Feedback. Practical alternatives to feedback widgets, survey prompts and NPS emails ↓ ✅ Feedback widgets help teams gather feedback at scale. ✅ E.g. in-app surveys, pop-ups, hints, panels, sidebar buttons. ✅ In-app surveys typically have response rates of 15-30%. 🤔 Side “feedback” buttons often have very low response rates. 🤔 Pop-ups perform better, but always (!) come at a wrong time. ✅ Ask for feedback only once a customer succeeded or failed. ✅ E.g. confirmation pages, finished transaction, failed import. ✅ Succeed → ask for improvements, user sentiment, slowdowns. ✅ Failures → ask to submit a problem, leave email, get a reward. 🚫 Instead of NPS, ask: “How easy was it to complete your task?” 🚫 Generic widgets → low response rates, low quality feedback. ✅ Ask specific questions about a process a user just finished. ✅ Ask to choose 5 tags (e.g. good price, noisy) that describe UX. ✅ Suggest to add details, tags, severity level, images, screenshots. ✅ Start with closed-ended questions (ratings, multiple-choice). Personally, I find in-person usability sessions infinitely more insightful as we can see people completing tasks and ask for feedback directly. But: it doesn’t work at scale. On the other hand, general feedback, such as NPS score, is rarely actionable, and simple thumbs up/down aren’t very insightful. And it’s very difficult to get actionable feedback when users experience severe failures. The best way to gather feedback is to ask for help and to be helpful. Ask for very specific feedback about a very specific feature that a user has just interacted with. Suggest a way out when a user experienced failures or failed transactions. Tailor questions to their context. For example, on a product page, ask about product clarity or ease of finding information there. But keep in mind that users who engage with feedback widgets may not represent the entire user base. First visitors are unlikely to provide any meaningful feedback, and annoyed customers often exaggerate their troubles. And depending on when and how the feedback is asked, the outcome can be remarkably biased, flawed and drive to wrong conclusions. So don’t draw big conclusions from surveys alone. Getting them right is very challenging, so whenever possible, complement them with user interviews, observations and UX research. Useful resources: UX Guidelines For User-Feedback Requests, by Anna Kaley https://lnkd.in/ehK3bAKm How To Set Up In-App Feedback, by Daniela Nguyen Trong https://lnkd.in/e3pByE5W A Guide to In-app Surveys for SaaS, by Moritz Dausinger Guide: https://lnkd.in/ertVRKmv Data: https://lnkd.in/eAJHK7tT Microsurveys Database (Notion) https://lnkd.in/e9-hAH7Z #ux #research

🚀 Smart Interface Design Patterns (https://lnkd.in/d4CNaTxR), a practical, growing video library on UX that I’ve been working on for the last 4 years — with 100s of practical examples, from complex data tables to configurators to dashboards UX (coming up this year!). Don’t forget to use the code LINKEDIN to get 15% off. 🍎 Best practices, examples and UX guidelines. 🍿 No subscription: get once, access forever. 🍰 5 new video chapters added every year. 🍧 Add-on: live UX training, next in Nov 2024. 🍒 Friendly team options for great teams. 🍭 Use the code LINKEDIN to get 15% off today. Free lesson: Designing Search UX in 2024 (2.5h) https://lnkd.in/dc-pWS-m Free lesson: Usability vs. Security https://lnkd.in/d6YNf9TJ Free lesson: Language Selector UX https://lnkd.in/dSfGhJDS All lessons (10h, with 35 video chapters) https://lnkd.in/d4CNaTxR Convince Your Manager (PDF) https://lnkd.in/eDD8YD78 I’ve tried my best to keep the pricing for the videos and the UX training affordable for everyone. But if it’s still a bit too expensive, please send me a message and I’ll do my best to make it work — for you and for your incredible team. 🏕️ Friendly Team Packages If you’d like to get a friendly package for your team, please send me a message and I’ll do my best to make it work. No subscription: your team will get life-time access to all videos, with 5 new videos added every year, for free. 🌎 Parity Pricing: Send A Message! I’ve tried my best to keep the pricing affordable for everyone. But if it’s still a bit too expensive, please do send me a message and I’ll do my best to make it work — for you and for your incredible team, of course! 🎉🥳 🚀 Convince Your Manager PDF A friendly guide to help you and your team get a buy-in for the video library and the UX training, with all the fine details of what it provides and how it works. https://lnkd.in/eDD8YD78 If you know a friend, neighbour or total stranger, please share the word — in Slack, Teams and casual conversations — if possible. And thank you for your incredibly kind and warm support, everyone! I sincerely hope you’ll find the course helpful for your work! 🙏 🙏🏼 🙏🏿

🎡 UX Workshops With Users (Scripts + Templates) (https://lnkd.in/evqDZSFe), a helpful overview of practical techniques to turn a verbal-only interview into a collaborative UX workshop — with sticky note mapping, card sorting, feature location, solution drag’n’drop and voting. Kindly put together by Laura Eiche. 👏🏽 🤔 Users and designers often a speak a different language. ✅ Insights are clearer when you see users performing tasks. ✅ Switch question-answer sections with small visual tasks. ✅ Sticky note mapping: for user flows, journeys, org maps. ✅ Card sorting: organize data, filters, menu items into groups. ✅ Feature location: ask users where they’d expect a new feature. ✅ Drag’n’drop: ask users to design their own UI or page layout. ✅ Solution voting: get feedback on many design directions. ✅ When explaining a task, show what you’d like them to do. ✅ Track where users are undecided, and follow up in a debrief. Personally, when I jump in a new project, I like to run walkthroughs with actual users as a way to understand the domain and the product. I invite them to show me the product, how it works, the features they use, the quirks they’ve discovered and the shortcuts and loopholes they rely on. And then I ask them to actually complete a few routine tasks on spot. From start to finish. That’s when insights emerge, and that’s when you might notice that the things said and the things done are not necessarily the same thing. And: you might not need to ask users to speak loudly about their experience. No ongoing conversation, no answers, no confirmation — keep it as realistic as possible, and in the setting that the customers are actually used to. 🗃️ Useful resources: UX Research Templates, by Odette Jansen https://lnkd.in/ekMDU9yJ Usability Testing Notion Templates (Script + Report), by Slava Shestopalov https://lnkd.in/gZyBtK6u Remote UX Workshop Activities, by Jordan Bowman https://lnkd.in/e8wSMVwC UX Workshop Scripts + Templates https://theuxcookbook.com UX Research Field Guide (Books + Templates), by UserInterviews https://lnkd.in/e4Ygsyuu Complete Guide to UX Interviews (+ Template), by Anna Savranska https://lnkd.in/eKt2WmrJ --- 🧲 Miro and Notion templates: UX Research Templates (Miro), by ServiceNow https://lnkd.in/e48nKzKA User Task Canvas (Miro), by Geoffrey Crofte https://lnkd.in/eFaD3Rmu Miro Templates For UX Designers https://lnkd.in/e8Hkp-ws Notion Templates For UX Designers https://lnkd.in/en_VBc6r #ux #research

Design Systems Playbooks And Helpers (https://lnkd.in/eKwNGaaR), a wonderful growing repository of helpful Figma playbooks and templates for designers who work on design systems — from strategy map and audit template to tokens naming, metrics and design documentation. Neatly put together by Romina Kavcic (highly recommended to follow!). Multi-Brand Design System (attached): https://lnkd.in/e2i92PQK Design Tokens Naming Playbook https://lnkd.in/eSjc59CG Design System Strategy Map https://lnkd.in/eZXuYXDz Design Systems Checklist https://lnkd.in/e5sCAJks Design System Metrics Collection https://lnkd.in/eSSJPYrb Design Documentation Template https://lnkd.in/eR89hEZd Design Audit Template https://lnkd.in/emAQCbdp 🧲 Useful resources: Design System In 90 Days Canvas (FigJam), by Dan Mall https://lnkd.in/evtWqQnv Design System Checklist Workshop (PDF), by Nathan Curtis https://lnkd.in/eTUusGcw Design System Canvas (FigJam, PDF), by Marianne Ashton-Booth https://lnkd.in/euRuGVri Multi-Brand Design System Figma Kit, by Pavel Kiselev https://lnkd.in/eShgnPnW Vodafone Variables Taxonomy Map, by Alexandra Dyulgerova, Robyn Layton https://lnkd.in/eYi-GQwp DoctoLib Design System Organization, by Jérôme Benoit https://lnkd.in/e9343uqv Figma Variables In Design Systems, by Nana Chon 🧙🏻♀️ https://lnkd.in/eKNZsd2H How To Launch Design System Components (+ Figma Template), by Rama Krushna Behera https://lnkd.in/ezPq2hQp Design System Process Flowcharts, by yours truly https://lnkd.in/eHyP65EM And a HUGE thank you to Romina and all the wonderful contributors above for putting all these guides and helpers together, and sharing them so kindly and passionately with the entire community. Your efforts do not go unnoticed. 👏🏼👏🏽👏🏾 Please support them if possible — they usually have wonderful books, guides, courses and trainings that are worth looking at, and that your team could benefit from. 🙏🏾 #ux #designsystems