Learning from Giants #39
Learning from Giants #39
Visual design rules you can safely follow every time, Marc Andreessen's take on the only thing that matters for startups, and how Spotify optimized their largest data transformation flow ever.
👋 Hi, this is Mathias with your weekly drop of the 1% best, most actionable, and timeless resources to grow as an engineering or product leader. Handpicked from the best authors and companies.
Did a friend send this to you? Subscribe to get these weekly drops directly in your inbox. Read the archive for even more great content. Also: I share these articles daily on LinkedIn.
Visual design rules you can safely follow every time
You've probably had to give feedback on UI or build UI yourself in the past. And what seemed like an easy task became daunting with the addition of design.
You can tell some UI doesn't look professional. You can feel it. But are unable to explain it. Similarly, when building it, you need outside help to produce something satisfactory. You're lacking something essential: design rules. And here's where to start:
"You do not have to follow these rules every time. If you have a good reason to break any of them, do. But they are safe to follow every time."
First, a core principle:
"Everything in your design should be deliberate.[...] This means whitespace, alignment, size, spacing, colour, shadows. [...] If you do not do this your design will not feel coherent."
📗 Anthony Hobday's Visual design rules you can safely follow every time is a collection of rules for your next UI review or building session. If you're in product or software, it's a great way to start thinking about design, and being intentional about it.
Here are a few examples:
"Use near-black and near-white instead of pure black and white. Pure black looks unnatural on a screen, and pure white is too bright."
"Lower letter spacing and line height with larger text. Raise them with smaller text."
"Make horizontal padding twice the vertical padding in buttons."
But the article has plenty more!
The only thing that matters (is your market)
"This post is all about the only thing that matters for a new startup." Marc Andreessen, 2007.
In this post, Andreessen explores one of the most discussed startup questions.
"What correlates the most to success—team, product, or market?"
"The caliber of a startup team can be defined as the suitability of the CEO, senior staff, engineers, and other key staff relative to the opportunity in front of them."
"The quality of a startup's product can be defined as how impressive the product is to one customer or user who actually uses it."
"The size of a startup's market is the the number, and growth rate, of those customers or users for that product."
If you read startup and VC media, it'll always be about team and product. ProductHunt does not have a "biggest market" award, but they showcase the product and the team. Ask VCs and most will say the team is what matters.
"Personally, I'll take the third position—I'll assert that market is the most important factor in a startup's success or failure. [...] In a great market—a market with lots of real potential customers—the market pulls product out of the startup."
So what's that only thing that matters, Marc?
"The only thing that matters is getting to product/market fit."
📗Marc Andreessen's "The only thing that matters" is a timeless post from one of the world's most successful entrepreneurs and VCs. Of course, there are counter-examples of teams and products so innovative that they create their own market, but that's a tiny minority of startups, and all of them still had to get to PMF!
How Spotify Optimized the Largest Dataflow Job Ever for Wrapped 2020
"Sort Merge Bucket is an optimization that reduces shuffle by doing work upfront on the producer side. The intuition is that for datasets commonly and frequently joined on a known key... we can write them in bucket files with records bucketed and sorted by that key."
By writing datasets in bucket files with records bucketed and sorted by a known key, the Spotify user id, the disk, and network I/O of moving key-value pairs around is reduced to within that bucket. This technique effectively partitions most large-scale jobs into smaller-scale independent jobs, given that they do user-scoped transformations.
"Over the last year and a half, we've been adopting SMB at Spotify for various use cases and accumulated many improvements to handle the scale and complexity of our data pipelines."
Wrapped 2020 is Spotify's generated per-user recap of what they've listened to in the past year. SMB enabled the team to save a lot when processing a year's worth of data, because it was already partitioned by user.
"We estimate around a 50% decrease in Dataflow costs this year compared to previous years’ Bigtable-based approach."
📗 Neville Li's How Spotify Optimized the Largest Dataflow Job Ever for Wrapped 2020 is an interesting read because it shows that even in the era of big data, being intentional about how the data is stored can lead to significant optimization and cost savings.