TMW #095 | Headless Everything
Welcome to The Martech Weekly, where every week I review some of the most interesting ideas, research, and latest news. I look to where the industry is going and what you should be paying attention to.
👋 Get TMW every Sunday
TMW is the fastest and easiest way to stay ahead of the Martech industry. Sign up to get the full version delivered every Sunday for this and every TMW, along with an invite to the TMW community. Learn more here.
Here’s the week in Martech:
- Headless everything: Martech and the data warehouse
- Netflix’s ad business: Underwhelming, no innovation but a strong branding proposition
- Slowing Martech spending: Have we passed the pandemic golden age?
- Everything else: AI creativity, the 2nd age of Martech, peak shopping season, B2B marketplaces, data lineages, and rappers and their Bored Apes
Will everything in Martech eventually become a bolt-on to a data warehouse? The question matters insofar as you think about the technology needs of the marketer and the role of SaaS companies. If Snowflake had its way, the answer is probably yes.
With the advent of SaaS, companies have been managing many “mini databases” across many disparate systems. For every software company you’re buying, you are either creating new or storing old data in a system that the software business owns and manages. And in most cases, you never actually see the database that it’s stored in. Is it hosted on AWS? Or Google Cloud? Who cares? If the app works as I want it to, I don’t need to know.
Packaged and sold software in companies is now a $900 billion dollar industry. We are buying software companies to do everything from HR to accounting to email marketing to organizing office snacks. And with each product comes more data to manage.
But what’s also growing is the adoption of the cloud. Okta ranks AWS and Google Cloud as the platform's two most popular software companies with 14% of users implementing multi-cloud infrastructures. Gartner points out that total spending on public cloud solutions has grown 27% last year.
What’s important is not the way the data is stored but having confidence that you own and can control the data. But how did our thinking shift in this way? If data is an asset, shouldn’t we control the means of managing, transferring, and analyzing that data?
This is the question data and analytics engineering teams ask themselves often. And the tide is turning towards the data warehouse in the cloud as the best way to centralize all the data stored in disparate apps. And because of this Martech is increasingly becoming more “headless” – solutions that bolt on top of pre-existing databases.
The concept of “headless” is often used in the context of the CMS – a way to decouple content storage with content delivery. But the concept should also be applied to many other areas of Martech as delivery is decoupled from storage in different ways.
As one example, Darrell Alfonso, the global marketing operations manager at Amazon, explains in an episode of MSoM, that there is a need for decoupling marketing automation platforms from their own databases. Data hygiene and privacy practices are often not followed where marketing automation platforms store data, and now, more than ever, marketers are needing transactional, product, and customer service data to better tailor communications. Doing this is more effective when using a platform that doesn’t store data in its own right, but sits on top of a data warehouse. Darrell suggests:
"Every marketing automation platform will eventually not have its own way to store data, but rather will sit on an existing database too. If we look at the amount of investment into cloud architecture this seems to make more sense. If all we really need is an API layer, with tools to activate data, send it to a BI tool or create customer experiences directly on top of it, does that make more sense? I think it does."
The accessibility paradox
The problem with making everything in Martech headless is the accessibility paradox. As much as I like the idea of having marketers, products, and digital folks having the tools at arm’s reach to ship their own ideas, expanding that kind of access to core systems like cloud-based data warehouses increases risk.
However, If tools built on top of the warehouse require skills and knowledge of database management or SQL then we’ve come full circle in SaaS needing to employ engineers and developers to make modern marketing work in the enterprise.
Don’t get me wrong, there are ways around this paradox. One of the ways is to compartmentalize software for various teams. One example is Castled Data which has built a marketing automation platform that is plugged directly into Snowflake or AWS. They delineate between the data engineering team that set up the integration and schemas and the marketers who will create segments and activate them across channels.
Another way to enable the flexibility of a headless solution, but also enable non-technical people to use it is Plasmic which allows you to use No Code tools to build websites and apps on the code and databases that already exist. Softr, is a No Code app that does something similar, but with managed databases like Airtable. Looker, the dashboard software company, also runs along these lines – let your engineers set up your schemas and integrations, and marketers can create charts on top of it. By decoupling the data layer from the execution layer, various teams can work on what they need to work on.
Costs to compute
There are some assumptions in the idea that everyone would want this kind of setup. One is the kinds of data that sit in a data warehouse. For example, to pull everything into AWS for a specific point solution to do its job, say for example web browsing data, for large companies you’re talking billions of events streamed through the platform. That’s unsustainable if you’re just after a web performance dashboard.
Besides, using a cloud platform for everything comes at a cost. As Stas Sajin in his piece, Why is Snowflake so expensive? suggests the compute and storage costs of managing data in and out of a warehouse can add up exponentially and create unsustainable situations for brands. Currently, a Snowflake customer that joins a year ago is spending $1, and now they are spending $1.7 a year later. And because of this, the company has increased revenue 105% since 2021.
In some ways, going headless makes a lot of sense, in other ways it doesn’t. But as companies continue to build greater sophistication and skills with the cloud, I can only expect to see more startups bolting on pre-existing solutions onto data warehouses. This is good news for Snowflake, Amazon and Google Cloud Platform. But is it good for the marketers that are trying to turn customer data into something valuable? Time will tell.
📈Chart Of The Week
Slowing Martech spending. According to eMarketer, Martech is set to surpass $6 billion in spending this year. Which is a significant decrease in year-on-year growth compared to the past two years. Link
📰 Latest Developments
Netflix and ads. More information about the ad network for the streaming platform has been released. Not much in the way of targeting options. But there’s a lot of value in the branded positioning with Netflix content anyway. This is shaping up to really be more like Cable 2.0 than anything groundbreaking or innovative. NEWS. ANALYSIS.
AI creativity is exploding. Stable Diffusion is the latest app to launch over the past two weeks that does a similar job to DALL E 2 and Midjourney with GAN-based illustration and design. Only this time, Stable Diffusion has been made available to the general public instantly (not like the DALL E 2 waitlist approach). The one interesting feature is the app’s ability to use visual prompts (like a crudely created visual prompt, or a partial image) as the basis for its artwork. AI in the creative services industries is growing faster than we can manage its implications on intellectual property. STABLE DIFFUSION. ANALYSIS.
Procter and Gamble support UID 2.0. One of the world’s largest advertisers is working with the Trade Desk on using the tracking and targeting solution. The challenge with CPG brands like P&G is the lack of first-party data to make UID 2.0 work. The other challenge is having a customer identity solution roll out across publishers that is also owned by one company. If UID 2.0 become mainstream, this might actually introduce centralization problems. If it doesn’t then publishers will have trouble managing dozens of identity tags on their websites. Link
TikTok as a search engine. A deep dive into why young people default to TikTok for search and the implications for marketing, discoverability of products and services, and misinformation. (Pairs well with TMW #088 and #094). Link
The 2nd age of Martech. Scott Brinker on the trend towards ecosystems, blended software and services, and customized apps on commercial platforms. The cloud is still early, and Martech is still converging on it. Link
Does being data-driven lead to good decisions? Being data-driven has become a mental shortcut for marketers to justify a way to come to conclusions. Two articles discuss the need for more reasoned thought over and above “following the data.” INDUSTRY. ANALYSIS.
🔢 Data & Insights
Retail media. Walmart made $2 billion in ad revenue last year. Link
Child safety. 64% of children’s apps on the Apple and Google Play stores access personal data and 96% transmit that data to ad bidding networks and Adtech data brokers. Link
An analysis of this year’s peak shopping season. Less spending generally than last year’s shopping season. And a big decline in eCommerce. Link
Is there a market for B2B marketplaces? An analysis into what the author suggests is a $1 Trillion economy waiting to happen. Marketplaces have been focused on consumer-facing use cases, but what do some of the biggest marketplaces do for business buyers and sellers? Link
Creating a lineage for data. Here's an approach to building a map of where all the data lives in an organization. The challenge is doing it so that everyone can understand it. Link
DALL E 2 prompt marketplace. Cheaper than traditional artwork stores, and faster than hiring a freelancer. Upwork, Fiver, and Shutterstock - it was good knowin’ ya. Link
✨ Weird and Wonderful
Snoop, Eminem, and the Bored Apes. A lot of criticism of a new music video featuring the two rappers using their BAYC intellectual property. Most of the online discussion about the video (outside of dunking on the stylistic choices) is that the video doesn’t do enough to onboard users to Web3. Link
Instagram verification schemes. This kind of thing has been going on for years. Since when did a verified blue check become such a force of social status? Enough to shell out a lot of money to get one? Link
Making money from kids yelling “poop” at Alexa. On the internet, there’s an economy for everything. Link
Make sense of marketing technology.
Sign up now to get TMW delivered to your inbox every Sunday evening plus an invite to the slack community.
Want to share something interesting or be featured in The Martech Weekly? Drop me a line at firstname.lastname@example.org.