Rohit Murarka’s Post

Do a Wrapped for Your Company Spotify started this "Wrapped" trend which is essentially showing you all the great "things" you did in 2024. We are pulling together our Baltoro Wrapped data. I highly recommend you do this for your company or team. 1 - It is great to see and show your team what they accomplished 2 - You will see some patterns of wins that you can take into 2025 #wrapped #companywrapped

Lost in Backstage Updates? Let Spotify's Maintainers Be Your Guide! https://lnkd.in/dJ9bEehf Is your Backstage instance stuck in 2023? Discover how the latest updates can supercharge your developer portal! 🚀 📽️In this video, Ben Lambert and Patrik Oldsberg from Spotify reveal: ✨The new Backend System ready for production ✨Exciting Declarative Integration features ✨The Community Plugins repository for easier contributions ✨A peek at upcoming Dynamic Features Watch the video to see how these Backstage updates can become your dev team's roadmap to enhanced productivity. P.S. Already using Backstage? Share your upgrade experiences in the comments! 💬 #lowops #backstage #DeveloperPortal #PlatformEngineering #OpenSource

Spotify quietly phasing out Car Mode piqued my interest as a PMM. If I were at Spotify, here’s how I’d use Enterpret to manage the Car Mode update: 1. Find Users: Create a group in Amplitude of all the users who used Car Mode. 2. Connect Quant <> Qual: Use our Amplitude integration to tie product data to user feedback in Enterpret. 3. Dig into Feedback: Identify key themes using Enterpret and use those keywords to find different user groups, like truck drivers, Uber drivers, or carpoolers. 4. Tailor Messages: Use these insights to craft personalized messages for different user segments. 5. Close the loop: Share these user groups and messaging ideas with the marketing team to proactively keep customers informed. For extra credit, I'd use Enterpret's Synced Users and Accounts to pinpoint users by account type, and then have messages specifically for Premium users. 😁

It's time for our latest installment of testing Enterpret models on public data 👩💻today, we're putting the spotlight on Spotify! Looking at public feedback, it is indisputable that Spotify users carry a deep love for the music-streaming player 🫶 As we dug in, an 🚨Anomaly🚨 from April and May popped up about Spotify removing Car Mode. (Our Anomaly detector identifies statistically significant increases in volume based on previous time periods to help companies pinpoint surging topics.) We looked at the history of Spotify's driving features, from Car View to Car Thing and surfaced why users want Car Mode back! Click through to learn what we found, or get the link to the full post "Spotify's Driving Features: A Bump in the Roadmap" in the comments.

Here's our slightly more corporate take on the popular trend...Customer Impact Wrapped 2024! 🚀 Spotify's 'Wrapped' is a genius way to keep customers engaged and excited year after year! By tapping into personal data and showing users their unique listening habits, Spotify creates a sense of nostalgia and pride, making it a fun and shareable experience. #SpotifyWrapped #CustomerLoyalty #EngagementTactics #CustomerRetention #BrandLoyalty #MarketingStrategy

Attention is to Big Tech what land was to feudal lords. When we’re enclosed within the walls of a digital fief, neither the old-fashioned commodity-producing capitalists nor the masters of manipulation they hire to advertise their wares can reach us. Unless the lords of these fiefs open the gates and let them in. Which, of course, they do – for a fee. So far, so feudal. What about the “techno” part? Well, once it has our attention, Big Tech can modify our behavior. Its algorithms first build trust. Spotify diligently learns what we like; Alexa helps us with chores. Then they nudge us to give up more information so they can cater to more of our whims. We’re training them to train us to train them. This cycle locks us into a fief. Even if Spotify’s sound quality is worse than its rivals’, users mostly won’t make the switch – those algorithms just don’t know them like Spotify’s does, after all. Copied - TechnoFeudalism

Every company hates Kubernetes disasters when they are at the center of it. But DAMM does it make for a good war story 👨🏼🚒 Spotify deleted their production clusters, and the best part is that end users didn't experience any significant service disruptions 😯 Check out this great KubeCon talk about how they accidentally deleted all their clusters across multiple regions, and yet users barely noticed 👇 https://lnkd.in/dHgVukNN

Spotify Lesson 2: If I could have seen my Spotify stats through the year, would it have changed what I listened to? Like, if I'd known how embarrassed I'd be publicly post I'd listened to Mean Old Sun more than other song in 2024, I'd probably have put something else on repeat to push it further down the list. Does it matter? Actually, sometimes it does. What gets measured gets managed. Once upon a time, I was on a project and our KPI was the number of times someone logged into a system. Of course, we wanted them to 𝘢𝘤𝘵𝘶𝘢𝘭𝘭𝘺 𝘶𝘴𝘦 the system, but instead they logged in, then logged right out and didn't use it. We weren't getting feedback, so we thought things were working well... because no-one was using it, even though our KPI showed they were. Measuring the right thing is a critical part of data science and analytics, but not one we think about or talk about as often as we should. Would knowing your Spotify stats have changed what you listened to through the year?

🎶 It’s Spotify Wrapped season! Ever wonder how the magic behind Wrapped *actually* happens? Spoiler: It's not really magic- it's data! Check out this article by Spotify’s engineering team to learn how they optimized one of the largest dataflow jobs ever to make Wrapped possible: ⬇️ https://lnkd.in/e9MiAgQu What’s on your Wrapped this year? 🤳

Trigger analysis has been gaining well-deserved attention recently, both for ML evaluation and A/B testing in general. At Spotify, we have been using trigger analysis, also known as exposure filtering, for many years. While it is a powerful technique to enhance efficiency, it does come with certain drawbacks and pitfalls. In this post, we discuss the trade-offs and share our learnings.