How Adobe.com Uses I/O Runtime to Optimize On-Site B2B Personalization

Effective personalization of digital experiences is more than a ‘nice-to-have’ for websites and apps; it is now an expectation of users that, when they go to a company’s website, it is relevant for them and their specific needs. Meanwhile, marketers want to make use of the latest technology in optimizing experiences for each user. This presents a big resource challenge for companies, even ones the size of Adobe; how do you make sure your website and CRM data align, without needing to manually keep all this data in sync?

Adobe I/O Runtime is Adobe’s serverless platform; it allows developers to run code on top of Adobe’s cloud solutions, bringing multiple services together with APIs. For the team that runs Adobe.com, I/O Runtime presented a new opportunity to deliver more personalized experiences, augmenting their existing workflow with Adobe Audience Manager.

“We wanted to access CRM attributes in Audience Manager, as part of our desire to enable Account Based Personalization for our B2B audience. Although we’re able to access the Company and basic firmographic information of the Adobe.com visitor with the help of a third party, the richer CRM attributes that we have in our databases, such as identifying existing customer accounts, industry classification, or product ownership, were not readily accessible on Adobe.com,” explains Fred Kuu, who leads a team of solution architects on Adobe.com.

“We want to gain access to Account-level attributes for use in Audience segmentation in Audience Manager. These attributes are defined in our CRM systems, but not directly available on Adobe.com — example attributes like identifying existing customer accounts and product ownership,” said Fred Kuu.

“The previous approach required us to query our database, create a classification file classifying Company Names with their respective attributes, and then in Adobe Audience Manager, manually create the derived signals. This process could take days. When there was an update required, we’d need to delete the derived signals and repeat the process.”

The Adobe.com team’s previous approach to audience segmentation included a lot of manual data entry.

Fred and his team are responsible for developing new digital capabilities for Adobe.com. When he and his team heard about I/O Runtime, they did some investigating to determine how it could possibly help them deliver more effective personalized experiences to customers. “One of our architects quickly honed in on the ability to implement real-time data lookups or queries as part of Runtime function. He began exploring how best to access the CRM data. Getting access to CRM attributes in real-time and sharing to Adobe.com would solve what has traditionally been one of the biggest pain points in our Personalization efforts,” said Fred.

How Adobe.com implemented a more effective process of personalization

Fred’s developer decided to create a function that would receive the Company Name as input, do a lookup in real-time, and then return the associated Attributes for that Company. These attributes would then be made available to Adobe Audience Manager and be used to build Audience segments. Since the Adobe.com team already had the required mapping data, all they had to do was develop the code to integrate the website with and do the lookup.

The Adobe.com team’s new approach to audience segmentation, using Adobe I/O Runtime, saves considerable time and resources.

“What started as an idea became code within a week and quickly launched on Adobe.com as a proof-of-concept. Since early December, the function has been running on Adobe.com. The developer was able to quickly grasp the concepts and go from whiteboard to this proof-of-concept in just one week,” said Fred.

“I believe we’re only scratching the surface in terms of what’s possible with Serverless”

As a result, creating Audience segments based on CRM data took much less time and manual effort; with I/O Runtime, it can be done in minutes. Activating new attributes or refreshing data would simply require refreshing the classification table within the Runtime function. Adobe is now working to leverage the segments in upcoming Personalization activities on the website and Media Retargeting.

While Fred feels this way of harnessing I/O Runtime has improved marketing efforts on Adobe.com, he feels the best is yet to come for Serverless solutions.

“I believe we’re only scratching the surface in terms of what’s possible with Serverless. Certain systems integrations that would’ve previously required significant custom development effort and operational maintenance becomes much simpler in a Serverless environment,” he said.

“Instead of worrying about complex integration builds and releases, teams can focus on functionality and developing simple functions and microservices that can be plugged in. This is significant for marketing organizations as there can be much more rapid development, including POCs.”


Fred will co-present a session about Adobe I/O Runtime, and how it can enrich data in near-real time to extend ABM personalization efforts, later this month. Join him to hear more about this and other I/O Runtime user cases at Adobe Summit, taking place in Las Vegas from March 26–28.

Follow the Adobe Tech Blog for more developer stories and resources, and check out Adobe I/O on Twitter for the latest news and developer products.