Beginning of the Endpoint: AWS, GCP, Snowflake, Databricks Compete/Connect in Marketing Cloud

The oil that makes the cloud run is data. At the start, AWS and S3 provided a simple storage service for data. It was web-based, secure, scalable and cheap. It changed the game. It’s been 17 years now since S3 launched and kicked off the “Cloud Wars.” Not only have Amazon/AWS Google/GCP and Microsoft/Azure been growing exponentially (and in the case of AWS & Azure printing money), they have been strategically positioning themselves horizontally and vertically. Bloomberg Intelligence analyst Anurag Rana recently estimated the value of AWS as a stand alone at $1.5 trillion to $2 trillion.

The rise of blogging, social media, apps & games happened alongside the maturation of the AWS. Almost every product and service on the web runs some part in not all of their systems on it.  Whole categories of AdTech and MarTech SaaS that couldn’t have been built without the cloud have already been through a few investment/M&A/IPO cycles. The result of cloud services and applications like Hadoop, Mongo, Redis, Elastic Map Reduce, EC2 and others that gained traction from 2008-2012 on advertising and marketing landscape looks something like this:

Today, after a decade of development and advances we sit with a new suite of cloud-based tools for marketing and advertising. The use cases may remain the same but the data processing, pipelines and tools are more advanced. They are easier to stand-up, scale, manage, monitor and are much more powerful, flexible and extensible.

The result of this data infrastructure maturation is this; we are moving from services and point-solution applications being built “on the cloud” and packaged as products you rent to services and point solution applications being built “in the cloud” and packaged as products you own. 

This is a sea change for brands, agencies, consultancies, investors and the technology marketplace serving advertising and marketing. Even the largest marketing companies like Adobe (Azure partnership) and Hubspot (?) are faced with company defining decisions on how to adapt their brand, products and pricing to work in the cloud market. Reality seems like the way the cloud has been used in marketing and advertising for the last 10 years is coming to an end.

Beginning of the Endpoint

Timing is everything and the rise of Cloud Marketing comes at the onset of increased needs for consumer data privacy. Now that data is being anonymized better than zero and with the advances in differential privacy and other data obfuscation methods we can start to move and join data in ways we couldn’t before. It’s worth noting to all the privacy skeptics that consumer privacy is an accelerator for the future of cloud marketing and with that advertiser and marketing performance. 

Cloud maturity means there is an increased awareness of data performance related to pipelines and a growing desire to derive intelligence and activations, often in real-time from this data. The architectural changes that have taken place over the past few years, namely the rise of data warehouses and data lakes are foundational and great for marketers.

We are at the beginning. But in the end all marketing will become Cloud Marketing. Cloud marketing is the natural progression from SaaS. We can see evidence of this with the rise of composable CDPs. I suspect this marketplace will look more like the App Store than enterprise sales forces fueled with VC. It will be a business of APIs instead of RFPs. It will be a business of selling enablement rather than solutions. Once the model has hardened it will likely be intractable for a century if not more. The only uncertainty is how long it will take to harden. But progress is accelerating and the Marketing Cloud as the endpoint is inevitable. 

I see three inflection points happening right now in Marketing Cloud that are having a large impact on this inevitability. In no particular order:

Google Analytics 4

Platforms controlled the destiny of the cloud mostly because with their data centers they built the cloud (though Facebook with its massive data center footprint focused on building goggles for its product rather than a cloud for its customers). A lot is written about Google being the dominant ad platform and ads subsidizing everything from wi-fi balloons to a Cloud business. What’s talked about less is Google’s utter dominance in web data creation, collection, analytics and activations

Google creates data for brands (Search/Display) and then captures (Google Analytics), activates (Google Tag Manager), analyzes (Looker), collaborates (Ads Data Hub), and stores (GCP). This is an amazingly valuable closed loop for Google and their customers.

Starting this July 1st (with a 7/1/24 deadline for Enterprise) Google is replacing its current version of Google Analytics with an entirely new technology called Google Analytics 4 (GA4) that is not backwards compatible. With 43% of the analytics market for F500 and more than 70% of the overall web analytics market the transition to GA4 is more than a big deal.  It is a sea change for analytics and measurement of the transactional web. 

Google can also now compete from a product standpoint with event based analytics systems that are much better products that created a new category of analytics from companies like Segment, Amplitude, Heap and Snowplow. But product parity is not the most important feature. The reason GA4 is a big deal for Cloud Marketing is that Google is tying its best features into its BigQuery cloud for free.

Connection to BigQuery is right in the GA4 UI. Want to manipulate your data in ways not available in GA4? BigQuery. Want to combine GA data with other data sources for analysis and reporting? BigQuery. Want access to unsampled raw event data? BigQuery. Want real-time analytics? BigQuery. Want data viz (Google Data Studio fka Looker)? Big Query. Need to connect to the Ads Data Hub clean room technology to share data in a “privacy compliant” manner? Big Query.

In the same way Google used Google Analytics to subsidize the growth of Search Ads they are using GA4 to subsidize the growth of Google Cloud Platform. It’s a brilliant strategy on the face of it and one that will have a profound effect on the future of Marketing Clouds. And it may just be the thing that brings GCP into profitability.

Market Validation of Snowflake & Databricks

As of this writing these two combined are worth about $75B. And while they have already built some incredible businesses and are a must-have in some key verticals they both still need high growth opportunities. Marketing and Advertising are two large opportunities. 

Snowflake of course is ⅔ of that combined value. It’s well positioned in FInancial Services and Healthcare Life Sciences where commercial effectiveness is easier to analyze. It’s making progress in Media and Retail where (per Wanamaker “Half my advertising spend is wasted; the trouble is, I don’t know which half.”) ad effectiveness and path to purchase are much harder to measure.

Snowflake is in a great position here. They have already done what many people thought was never possible, which is to have an upstart VC incubated cloud provider become a serious challenger brand in the cloud space. And they did it by solving business use cases with data warehouse as much as technical ones.

Granted Snowflake has serious work ahead of it in Media & Retail. It is currently behind in the clean room hype cycle, though this could prove advantageous in the long run. I am also extremely bullish on the connected application space as the way to solve the use cases in media and retail, simply because each sub-vertical and product/customer bases are so unique. Personalization, Customer Journey, Attribution and Media-Mix are all perfect for the connected app architecture Snowflake goes to market around.

Databricks is in a nice position too. Being a lakehouse rather than a warehouse they are cloud-agnostic. Because it was built on Apache Spark, Databricks is especially proficient for streaming data, low latency applications, ML/AI as well natively handling all types of unstructured data. With these features it makes sense for them to concentrate on media and marketing growth since these verticals move at the speed of the customer.

Attribution, as well as all types of scoring and matching decisioning/measurement play very well with Databricks and Databricks play very well with all everyone else in the cloud. That type of architecture is a good formula for success across the media walled gardens landscape which now also includes a growing number of walled gardens in Retail Media.

Over the past few years Snowplow and Databricks have emerged as serious players in the cloud. That their tech is well suited for Marketing & Advertising use cases puts them in a great position. They likely already have the hearts and minds of CTOs. How well they win the hearts and minds of CMOs may be their deciding success factor.

AWS/Amazon Marketing Cloud (AMC)

At re:Invent in December I asked an AWS exec about their CDP strategy with composable CDP emerging as a threat to the CDP vendor space. He said AWS focuses on customers and they are all AWS customers. It’s a very powerful statement of fact. Amazon has the largest stack at the table and is holding two aces.

Similar to Google, Amazon has a huge advertising business and is a market maker in the transactional web. It has the benefit of giant customer relationships and footprints in Cloud and Ads/Marketing including being the infrastructure for so much. 

Also similar to Google, Amazon is using its ad business to grow its Marketing Cloud. Recently Amazon launched AMC (Amazon Marketing Cloud) that integrates its DSP into a dedicated Clean Room environment in AWS at no cost. The goal here is to provide deeper levels of measurement, analytics, insights and ultimately data collaboration and activation. If you want examples of MarTech and AdTech convergence, Google and Amazon are exhibits No. 1 & 2.

One service Amazon lacks that Google has in spades is the behavioral data collection outside its walls. I think Amazon will acquire a company in this space. This would close the loops similar to Google mentioned above that are important to advertising revenue growth but will impact cloud revenue growth as well.

Similar to Google, Amazon’s second ace is the ability to leverage its 300M first-party consumer relationships to create its own “universal” Identity Service. It’s hard to ascertain if this will ever happen but maybe the better way to think about it is “why wouldn’t it happen?” 

Identity resolution is *the* major technical issue facing privacy compliant advertising and current solutions outside the platforms might never work. Amazon is arguably already dipping their toe here with Buy with Prime. Any more moves will have major repercussions in both media and cloud and strengthen Amazon’s place in each.


With so much marketing tech built on AWS its partner ecosystem is unrivaled. None of this legacy cloud wants to be disintermediated by new shiny objects. As the original composable cloud and currently printing money, AWS can walk and chew gum in all directions. 

One new direction is appealing directly to the use cases of brands and their agencies and consultants. AWS for Marketing and Advertising launched last year going to market with a clear roadmap focusing on these use case:

Identity Resolution (Matching & Modeling)

Data Collaboration (Clean Room)

Measurement (Attribution)

Retail (Personalization)

Clean Room is the most interesting play at the moment as it appears quite disruptive. AWS stated at re:Invent that you can spin up a Clean Room “in minutes” via their API. Snowflake has a Clean Room product but it is limited. Google has Ads Data Hub which nobody cares for and there is $100M in VC invested in Clean Room SaaS vendors. Clean Room can be a nice wedge for Marketing Cloud.

A sleeping giant for data sharing and enablement is AWS Marketplace. It has been around a long time but yet to find a fit in advertising and marketing. It might be a question of “when” rather than” if” for it to emerge as a force since conceivably it helps businesses move faster and data is a motor sport. For example, here’s a good looking KYC Segmentation Model with Realtime ML Based Inference nobody is using. Or did you know you can grab Twilio/Segment on there?

It will be fascinating to watch AWS and Amazon here. They are the leader but the rise of Snowplow and Databricks shows there are still market gaps. Of course those gaps probably get filled by AWS one way or another.

Layer Up

There are few surer bets than the cloud keeps growing and more and more business rules and decisions transact in and across these technologies. I highlighted a few things to watch but of course there are more. With use cases for data activations, audiences, content, conversion, customer journey, workflow and measurement (just to name a few), marketing shapes up as one of the most important areas for the Cloud over the next few years. 

That we are already familiar with terms like multi-cloud and hybrid-cloud is important too. It is after all the ability to interoperate and to build layers of services and applications that makes the cloud so appealing in the first place. This mix of competition and connection is great for customers. We have choices and can compose solutions that meet their needs both technically, computationally. For the marketing and advertising needs of businesses, full ownership of and sovereignty over data is the ultimate value prop. May the best performing data in the cloud win!





2 responses to “Beginning of the Endpoint: AWS, GCP, Snowflake, Databricks Compete/Connect in Marketing Cloud”

  1. Singularity: AdTech and MarTech are Converging Avatar

    […] On” a deep dive into real-time use cases and apps for streaming data. Then zoomed out on Marketing/Advertising GTM strategies of the Cloud Data Warehouses with “Beginning of the Endpoint.” This piece “Singularity” goes the last mile exploring the […]


  2. Goodbye Customer Data Platforms Avatar

    […] 2018 and I was blown away. 5 years ago AIQ was new and best in breed. However, that was long before the data warehouse and master data management became table stakes for serious businesses. It was […]


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Create a website or blog at

%d bloggers like this: