We Stitches share links among ourselves to pages we think our colleagues will enjoy. We think some of you will enjoy reading these too. Should we make a monthly roundup of curated links a regular blog feature? Tell us what you think in a comment.
The business of SaaS
From Stripe's blog we learned a lot about different models of SaaS business: low-touch ("some products sell themselves") and high-touch ("some customers need some help in deciding whether or how to adopt certain products"). The post goes into detail about the impact different models have on companies' growth paths.
Give meaning to 100 billion analytics events a day
Teads Engineering explained how it orchestrates Kafka, Dataflow, and BigQuery to ingest and transform a large stream of events. At Stitch, we process a large stream of events ourselves, and we just blogged the anatomy of our internal pipeline. Stitch and Teads have completely different use cases, but it's illuminating to see how others handle similar issues.
How we saved over $240K per year by replacing Mixpanel with BigQuery, Dataflow & Kubernetes
A post by the CTO of DoIT International gave us another look at a project that tackles big data. What do you do when you have so much data that your product analytics tool becomes too costly? That was the situation Jelly Button Games faced when its smartphone-based games became popular. To cut down the cost of what Jelly Button was paying for Mixpanel and build a more flexible data analytics pipeline, DoIT created a solution that involved ingesting both streaming and batch data into Google BigQuery using Google Cloud Dataflow. Lots of diagrams and code snippets in this piece provide details.
Top 8 Best Practices for High-Performance ETL Processing Using Amazon Redshift
Finally, we ran across a post on the AWS Big Data Blog that talks about optimizing ETL for Amazon Redshift. Hey, that's what we do! It was good to see Amazon's take on some of the lessons we've learned.
What interesting posts have caught your eye lately? Share them in a comment below.