DROP the modern data stack

Previously: the simple “modern data stack” when first discovered by data teams was magic. It made it absurdly easy to quickly centralize, transform, and present data for analytics
Now: the MDS is a landscape of hundreds of logos across all of “data” RIP MDS (magic data stack)

I read about every article on the Modern Data Stack (MDS) and listened to a few podcasts, youtube videos, and came to the conclusion: Everyone has their own definition of the Modern Data Stack

What does “modern data stack” mean?

Every week when I read about the modern data stack, it’s different. And it’s not a stack, it’s a collection of tools
- @oralassila @KGConferencekeynote TRUTH!!

I don't think anyone will be talking about modern data stack by the end of the year. The semantic satiation is slowly kicking in.

The main issue for me is what does it even mean?

What’s wrong with the modern data stack?

Today’s tooling makes data engineering easier. The tools of the modern data stack eliminate a lot of data engineering drudgery, and they enable us to do more than ever before with our data. So, what’s the problem?

Data engineering tweople: what do you consider as the "Modern Data Stack"? One particular set of technologies? A concept or an idea? The more I read about it, the more confused I am about the term.

Previously: the simple “modern data stack” when first discovered by data teams was magic. It made it absurdly easy to quickly centralize, transform, and present data for analytics
Now: the MDS is a landscape of hundreds of logos across all of “data”
RIP MDS (magic data stack)

“I read about every article on the Modern Data Stack (MDS) and listened to a few podcasts, youtube videos, and came to the conclusion: Everyone has their own definition of the Modern Data Stack”

What does “modern data stack” mean?

“Every week when I read about the modern data stack, it’s different. And it’s not a stack, it’s a collection of tools” - @oralassila @KGConferencekeynote TRUTH!!

I don't think anyone will be talking about modern data stack by the end of the year. The semantic satiation is slowly kicking in.

The main issue for me is what does it even mean?

What’s wrong with the modern data stack?

Data engineering tweople: what do you consider as the "Modern Data Stack"? One particular set of technologies? A concept or an idea? The more I read about it, the more confused I am about the term.

Previously: the simple “modern data stack” when first discovered by data teams was magic. It made it absurdly easy to quickly centralize, transform, and present data for analytics
Now: the MDS is a landscape of hundreds of logos across all of “data”
RIP MDS (magic data stack)

“I read about every article on the Modern Data Stack (MDS) and listened to a few podcasts, youtube videos, and came to the conclusion: Everyone has their own definition of the Modern Data Stack”

What does “modern data stack” mean?

“Every week when I read about the modern data stack, it’s different. And it’s not a stack, it’s a collection of tools” - @oralassila @KGConferencekeynote TRUTH!!

I don't think anyone will be talking about modern data stack by the end of the year. The semantic satiation is slowly kicking in.

The main issue for me is what does it even mean?

What’s wrong with the modern data stack?

Data engineering tweople: what do you consider as the "Modern Data Stack"? One particular set of technologies? A concept or an idea? The more I read about it, the more confused I am about the term.

Previously: the simple “modern data stack” when first discovered by data teams was magic. It made it absurdly easy to quickly centralize, transform, and present data for analytics
Now: the MDS is a landscape of hundreds of logos across all of “data”
RIP MDS (magic data stack)

“I read about every article on the Modern Data Stack (MDS) and listened to a few podcasts, youtube videos, and came to the conclusion: Everyone has their own definition of the Modern Data Stack”

What does “modern data stack” mean?

“Every week when I read about the modern data stack, it’s different. And it’s not a stack, it’s a collection of tools” - @oralassila @KGConferencekeynote TRUTH!!

I don't think anyone will be talking about modern data stack by the end of the year. The semantic satiation is slowly kicking in.

The main issue for me is what does it even mean?

What’s wrong with the modern data stack?

Data engineering tweople: what do you consider as the "Modern Data Stack"? One particular set of technologies? A concept or an idea? The more I read about it, the more confused I am about the term.

Previously: the simple “modern data stack” when first discovered by data teams was magic. It made it absurdly easy to quickly centralize, transform, and present data for analytics
Now: the MDS is a landscape of hundreds of logos across all of “data”
RIP MDS (magic data stack)

“I read about every article on the Modern Data Stack (MDS) and listened to a few podcasts, youtube videos, and came to the conclusion: Everyone has their own definition of the Modern Data Stack”

What does “modern data stack” mean?

“Every week when I read about the modern data stack, it’s different. And it’s not a stack, it’s a collection of tools” - @oralassila @KGConferencekeynote TRUTH!!

I don't think anyone will be talking about modern data stack by the end of the year. The semantic satiation is slowly kicking in.

The main issue for me is what does it even mean?

Previously: the simple “modern data stack” when first discovered by data teams was magic. It made it absurdly easy to quickly centralize, transform, and present data for analyticsNow: the MDS is a landscape of hundreds of logos across all of “data”RIP MDS (magic data stack)

@Sethrosen
Twitter

Previously: the simple “modern data stack” when first discovered by data teams was magic. It made it absurdly easy to quickly centralize, transform, and present data for analyticsNow: the MDS is a landscape of hundreds of logos across all of “data”RIP MDS (magic data stack)

@Sethrosen
Twitter

Previously: the simple “modern data stack” when first discovered by data teams was magic. It made it absurdly easy to quickly centralize, transform, and present data for analyticsNow: the MDS is a landscape of hundreds of logos across all of “data”RIP MDS (magic data stack)

@Sethrosen
Twitter

I read about every article on the Modern Data Stack (MDS) and listened to a few podcasts, youtube videos, and came to the conclusion: Everyone has their own definition of the Modern Data Stack

Doug Foo
Medium

What’s wrong with the modern data stack?

Today’s data tooling makes data engineering easier. The tools of the modern data stack eliminate a lot of data engineering drudgery, and they enable us to do more than ever before with our data. So, what’s the problem?

The data tooling explosion transformed a straightforward stack into a sprawling ecosystem of tools.

This ecosystem–which we’re still calling the modern data stack–is overwhelming. Then there’s the marketing. When the term “modern data stack” caught on, vendors (RudderStack included 😅) took full advantage. Now everyone is defining their own modern data stack confusing those who are actually trying to build one. With no single source of truth about the modern data stack and so many tools to choose from, data maturity can seem out of reach, but it’s more important than ever.

Let’s get real

It’s time to DROP the modern data stack and take a measured, more practical approach to these incredible data tools. It's not about a big step to implement the modern data stack. It's about understanding the small steps you can take today to make your stack more useful to your business and executing those steps in a way that supports progress for tomorrow. We call this the data maturity journey.

The Data Maturity Journey

The Starter Stack

Solve your customer data integration problems and send consistent data from your sites and apps to every tool through a unified data layer.

Diagram of data sources going to analytics and activation platforms.

Pain Points

Duplicative instrumentation and data flows send the ‘same’ data to different platforms in different formats
Inconsistent user profiles across downstream tools
Painful integration management requires constant attention from dev

Use cases to unlock

Streamlined data collection with a single SDK and API
Consistent user profiles in every tool, updated automatically with a single payload
Low maintenance data integration from a single, automated integrations layer

The Growth Stack

Unify your customer data in a centralized storage/compute layer and make every customer data point available for activation in every tool in your stack.

Pain Points

Inability to answer complex questions without brute-force data munging efforts
No complete customer profile or full view of the customer journey because your data is siloed in SaaS tools
Teams and tools at local maximums because valuable data in other tools isn’t available to use for optimization

Use cases to unlock

Central source of truth for all of your customer data
Complete customer profiles and a view of the customer journey that includes every data point
Data activation with data from the entire stack, so every team can drive optimization

The Machine Learning Stack

Move from deterministic to predictive analytics and calculate predictive user traits for use in downstream tools.

Pain Points

Optimization is limited because it requires you to discover both problems and opportunities before they arise
Data science can’t iterate quickly with fresh data because of outdated infrastructure and inconsistent or incomplete data
Syndicating ML model results downstream for activation is a major technical challenge

Use cases to unlock

More powerful features in your ML and data science modeling driven by a comprehensive data set
Clean, model-ready data without a massive data engineering lift
Data activation with data from the entire stack, so every team can drive optimization

The Real-Time
Stack

Deliver predictive, personalized experiences in real-time based on in-session user actions.

A diagram of data sources going into warehouse tolls through an in-memory database to an application layer.

Pain Points

Inability to deliver personalization in real-time using features from ML models
Inflexible personalization options from SaaS tools can’t handle complex use cases or integrate with your ML workflow
Sending in-session clickstream data in real-time is a major technical challenge

Use cases to unlock

Deliver in-session data in real-time to in-memory data stores with low-maintenance pipelines
Personalize every part of the customer experience using your ML models without the limits of SaaS
Automatically enhance your models with behavioral data from users who interact with a personalized experience

Where are you today?

All great journeys begin with a first step. Select your stage below to get your copy of our Data Maturity Guide.

Starter

Solving my data integration problems

Growth

Unifying my customer data

Machine Learning

Moving from deterministic to predictive analytics

Real- Time

Delivering personalized experiences in real-time