Castled.io

Excited to announce that Castled is partnering with Mitzu.io to create the first-ever warehouse-native retention and analytics engine for large B2C brands. 𝐓𝐡𝐞 𝐏𝐫𝐨𝐛𝐥𝐞𝐦 B2C brands store vast amounts of customer data in their cloud data warehouses, which they need to leverage to solve their retention and product analytics use cases. Currently, this process involves shuffling hundreds of millions of data points across your DWH, customer engagement tools, and product analytics platforms, leading to four major issues: 𝐃𝐚𝐭𝐚 𝐒𝐢𝐥𝐨𝐬: Traditional customer engagement and product analytics platforms operate in isolation, preventing a unified view of your customers for product and marketing teams. 𝐒𝐤𝐲𝐫𝐨𝐜𝐤𝐞𝐭𝐢𝐧𝐠 𝐂𝐨𝐬𝐭𝐬: These platforms charge based on the volume of data points ingested, causing costs to skyrocket as data consumption increases. 𝐋𝐢𝐦𝐢𝐭𝐞𝐝 𝐃𝐚𝐭𝐚 𝐔𝐬𝐞: Often, only a fraction of your customer data in DWH is used for retention and analytics purposes, leaving valuable customer insights untapped. 𝐋𝐞𝐚𝐤𝐲 𝐏𝐢𝐩𝐞𝐥𝐢𝐧𝐞𝐬: Data pipelines are often error-prone and cause delays, resulting in a suboptimal customer experience. 𝐓𝐡𝐞 𝐒𝐨𝐥𝐮𝐭𝐢𝐨𝐧: 𝐂𝐚𝐬𝐭𝐥𝐞𝐝 + 𝐌𝐢𝐭𝐳𝐮 With Castled + Mitzu, your customer data never leaves your data warehouse. 𝐔𝐬𝐞 𝐂𝐚𝐬𝐞: 1) Analyse user behaviour and create granular cohorts, such as churning users, in Mitzu.  1) Send these cohorts to Castled for engagement across multiple channels like email, SMS, WhatsApp, push notifications, and in-app messaging All without copying a single piece of data right where your data lives.

Excited to announce that Castled is partnering with Mitzu.io to create the first-ever warehouse-native retention and analytics engine for large B2C brands. 𝐓𝐡𝐞 𝐏𝐫𝐨𝐛𝐥𝐞𝐦 B2C brands store vast amounts of customer data in their cloud data warehouses, which they need to leverage to solve their retention and product analytics use cases. Currently, this process involves shuffling hundreds of millions of data points across your DWH, customer engagement tools, and product analytics platforms, leading to four major issues: 𝐃𝐚𝐭𝐚 𝐒𝐢𝐥𝐨𝐬: Traditional customer engagement and product analytics platforms operate in isolation, preventing a unified view of your customers for product and marketing teams. 𝐒𝐤𝐲𝐫𝐨𝐜𝐤𝐞𝐭𝐢𝐧𝐠 𝐂𝐨𝐬𝐭𝐬: These platforms charge based on the volume of data points ingested, causing costs to skyrocket as data consumption increases. 𝐋𝐢𝐦𝐢𝐭𝐞𝐝 𝐃𝐚𝐭𝐚 𝐔𝐬𝐞: Often, only a fraction of your customer data in DWH is used for retention and analytics purposes, leaving valuable customer insights untapped. 𝐋𝐞𝐚𝐤𝐲 𝐏𝐢𝐩𝐞𝐥𝐢𝐧𝐞𝐬: Data pipelines are often error-prone and cause delays, resulting in a suboptimal customer experience. 𝐓𝐡𝐞 𝐒𝐨𝐥𝐮𝐭𝐢𝐨𝐧: 𝐂𝐚𝐬𝐭𝐥𝐞𝐝 + 𝐌𝐢𝐭𝐳𝐮 With Castled + Mitzu, your customer data never leaves your data warehouse. 𝐔𝐬𝐞 𝐂𝐚𝐬𝐞: 1) Analyse user behaviour and create granular cohorts, such as churning users, in Mitzu.  1) Send these cohorts to Castled for engagement across multiple channels like email, SMS, WhatsApp, push notifications, and in-app messaging All without copying a single piece of data right where your data lives.

Super excited to announce our partnership with Mitzu.io — this is by far my favourite. I first met István Mészáros more than a year ago, introduced by a mutual friend. I was 100% bullish on the problem Istvan and his team were solving, as I had experienced it firsthand with traditional product analytics tools. We lost touch for a while since then, but Mitzu kept coming up in customer conversations — likely due to the similar warehouse native architecture. About a month ago, when I heard Mitzu mentioned twice in the same week, I knew I had to reach out again. This time, our perspectives had completely changed. We cared less about our warehouse native architecture and talked all about the restrictive customer experiences caused by existing data silos. And the customer experiences we can drive if we join hands. Consider this case: It's very common for a product team to create a cohort in a product analytics tool and send it to a CEP for engagement. Similarly, product teams often consume engagement actions from a CEP in an product analytics platform to unlock valuable campaign insights. However, for B2C brands, this means transferring hundreds of millions of data points every week, which is super expensive. More importantly, continuous delays and failures in these pipelines create major data inconsistency issues. Your product and marketing teams are NEVER EVER looking at the same customer view. With Castled + Mitzu, data never leaves your warehouse. Warehouse Native is the future — be it CDP, CEP, or Product Analytics, and it's a great time to be building one.

Today, we are announcing the general availability of our React Native SDKs. From the outset, we were clear that our broader ICP would be B2C companies with 1M+ users. And we were under the impression that companies at that stage would always invest in native technologies - Swift/ObjectiveC  for iOS and Kotlin/Java for Android. So, we refrained from supporting cross-platform technologies for quite a while and even rejected the increasing requests from our ICP for React Native and Flutter as "exceptions." About a year ago, we finally realised how wrong we were. For perspective, some of the world's largest mobile apps, like Facebook, Bloomberg, Walmart, Uber Eats, Discord, Skype, and others, utilise React Native under the hood. We launched the first version of the SDK last October. And after collaborating with customers and refining it over nearly 8 months, we are now excited to make it available to everyone. So, if you have your customer data in Snowflake, BQ, or Databricks and wish to use it to retain your customers on your React Native app, you know what to do!

Excited to share the first instalment of our series: "What They're Not Telling You" - 𝐇𝐢𝐠𝐡𝐭𝐨𝐮𝐜𝐡 𝐉𝐨𝐮𝐫𝐧𝐞𝐲𝐬! Please give it a read and share your thoughts 🙏

Is it founder bias, or does everything we read nowadays actually validate the notion that Zero Data Copy is the future? Just yesterday, Dell reported a data breach affecting roughly 49 million customers. The exposed data included names, physical addresses, and other product details. Should be a one-off, right? Nope. Microsoft, AT&T, and Dropbox are some of the other tech giants that have reported data breaches in the last 30 days, with sensitive customer information being leaked. If these heavy weights can be hacked, it's clear that no one is immune. Now, does adopting a zero copy approach to handling customer data guarantee protection against data breaches? For example, are systems like BigQuery, Snowflake, Databricks, or your in-house data lake impervious to attacks? The reality is, any system can be compromised. Nevertheless, businesses must collect and centralise customer data somewhere to operate effectively, and that place should ideally be your data warehouse. Choosing to maintain additional copies of your customer data outside your data warehouse is precisely that — a choice.

Technologies like Snowflake Unistore are ushering in a new era without redundant ETL pipelines. The key entity of Unistore, hybrid tables, can answer both transactional and analytical queries at the same time. But isn't this contrary to almost every fundamental database principle we have learned until now? I mean we have always used separate database classes to cater to unique access patterns — MySQL/DynamoDB for transactional workloads, Elastic Search for high-scale scans and data warehouses for analytical use cases. This separation made sense as it allowed database vendors to optimise the storage formats needed for specific workloads. So, how do hybrid tables achieve the same? The answer is surprisingly straightforward -- Hybrid Tables do not operate on a single copy of your customer data as you might think. Instead, it maintains two copies: one optimised for low-latency point lookups and another for data analysis. A data pipeline, operating under the hood keeps both the layers in sync. The smart query layer of Snowflake can then route the query to the appropriate datastore based on the actual query. We are possibly at the brink of an ideal world where all production and SaaS applications will be built on top of workloads like the Unistore.