Adobe Audience Manager (AAM) and Salesforce Audience Studio (Krux): Side by Side Comparison: Part II – Setup
In the first post of this series we discussed data collection. Whether it was real time date (Web/App) or onboarded data (CRM/Store), we highlighted how each of these platforms handles this important phase. It is essential that you are able to connect the pipelines fairly quickly and get data flowing, otherwise your low hanging fruit will stay hanging. Data collection is half the story when it comes to implementation, setup is where things get real.
I have sat in many sales pitches about the DMP and 90% of the time, the setup phase was skipped. Not on purpose but it seems to be of a lower importance when it was compared to data collection, use cases or activation. It is the reason why clients are surprised when it is taking so long to get value out of the DMP. No one properly planned for setup time.
In setup, extension of implementation, I will concentrate on the following:
Traits and Attributes
DMPs are not meant to collect everything and anything. They are not a very good analytics platform simply because they concentrate on the audience attributes and very little on time-based activities. This doesn’t mean that we don’t have the concept of time in DMP but it is limited. That being said, it is important to think through the taxonomy and data dictionary to bring into the DMP for two reasons: 1) Keep things clean and organized so you don’t overwhelm the power users and 2) Cut down the costs, you have a monthly quota of profiles or server calls that you want to stick to or otherwise the overcharges will creep in. Let’s discuss how each platform handles the signals flowing in:
Adobe Audience Manager
Given that Adobe already had an analytics platform, Adobe Analytics, they always championed the idea that you will have to define your data in the platform ahead of the data flowing into it. In other words, you need to setup traits (whether onboarded or real-time) before you open up the gates for data collection. Otherwise, your data will be ignored and lost, forever. You have the opportunity to see what data is being collected but ignored through a report, unused signals. Earlier this year, Adobe released Data Explorer (Signals) which is an extension of unused signals but on steroids: 1) You can see what you are missing and the volume of profiles associated with it 2) You can automatically create traits based on these unused signals and 3) You can retroactively retrieve this data and use it for traits and segments (this will come with a price though). For the traits that are accounted for, they tie to specific data sources, labeled and organized in a searchable folder structure and can be based on a single signal or many.
Salesforce Audience Studio
Since day one, Krux marketed their DMP that no data is ever lost. They called it “completeness” and the CEO, Tom Chavez, made sure to stress on its value in many of his keynotes, including this one. The idea is that any signal that comes into Krux, whether used or not, will be available and reusable at any point in time. This is extremely valuable especially when a new update to your site enables a new set of values that you might not have a use case for today but you will a month from now. Data in Krux is stored in the form of attributes which can be activated or not. Attributes have a display name, category and type but are not structured in any folder structure.
My POV
If this post was done 6 months ago, I would have happily said that Krux’s completeness is the winner here but the game has changed with Adobe’s release of Data Explorer. That being said, it is important to highlight few things:
-
Premium Fee
Data Explorer gives you the option to reclaim historical unused signals and turn them into traits which limits any data loss. Just be aware that this comes with a fee. There is no fee to see unused data and create traits to reclaim from that point on but to reclaim data retroactively is a premium option.
-
Folder and Taxonomy
Both tools have the option to categorize the traits/attributes but only Adobe offers a folder structure to store them. In Audience Studio, attributes are stored in a flat structure but are searchable with robust filters. This boils down to user preference on whether a folder or flat structure is better. I like the folder structure because it translates well from a taxonomy point of view especially if the DMP is to be leveraged by different agencies or groups within an organization. That being said, when you have many folders and subfolders, the Adobe UI becomes hard to navigate as you are limited to a fairly small space in the page’s real estate and have to depend on side bars and scrollers to navigate.
-
Completeness still a winner
Krux’s completeness is still a very powerful offering even with Adobe’s Data Explorer because it simplifies the number of traits/attributes you have to create. For example, let’s say you are an e-commerce business and have 1000+ products. In Adobe, you will need 1000+ traits to accommodate all the products while in Salesforce, you need one attribute. On the segment level, you can hook into all products from a list within that attribute. This is especially handy when it comes to capturing and building retargeting segments based on media campaigns.
Segmentation
Traits and attributes are one dimensional and tie to a single data source while segments span across multiple sources. Audience Segmentation is the biggest selling point for a DMP and both tools offer robust options to set them up. The comparison list for Segmentation can be a post on its own, in this section we will concentrate on the aspects that DMP gurus appreciate the most.
Adobe Audience Manager
Adobe offers one type of segmentation but gives you all the Boolean logic your heart might desire including recency and frequency. Segments are also stored in a folder structure that is tied to a taxonomy and audience strategy.
Salesforce Audience Studio
Salesforce split segments into three types: Standard, Realtime and Global Delivery.
Segments, similar to attributes, are categorized and stored in a flat non-folder structure. Segments also have the recency and frequency options across all 3 types.
My POV
In general, both offerings are very similar when it comes to segmentation but Salesforce goes the extra miles on few key features and functionalities that power users appreciate. I will highlight three here:
-
Device vs People Reporting
Audience Studio reports on segment uniques in devices (cookies, mobile ids) as well as people. Moreover, it breaks down the segment counts by device and browser (safari specific). This type of reporting is extremely valuable to know where this audience is reachable and to what extent they are impacted by ITP.
-
Referencing existing segments
This might sound silly but it happens all the time where you want to create a new segment based on an existing segment with a minor modification or addition. In Adobe, you can’t reference other segments or reuse existing segments in new segments but Salesforce gives you this option and it is beautiful. It takes away the pain of creating multiple different versions of a segment starting from scratch each time.
-
Forcing a user into segment
We will have a dedicated section on testing and troubleshooting in both platforms but I wanted to highlight this feature in Salesforce’s DMP. It is extremely easy to force someone into a segment, go into test mode and immediately verify whether the targeting experience is working properly or not.
Partners and Destinations
Once segments are created and populating the desired volumes, activation is the next logical step and both platforms excel in this space.
Adobe Audience Manager
Adobe supports different flavors of destinations: Cookie, Server to Server or Batch. AAM integrates with the top DSPs and Vendors in the space. In situations where you want to activate your segments with a new vendor, Adobe is willing to establish an integration with the vendor that can benefit you and other clients using that vendor. Destinations can be setup directly from the UI although the server to server ones are setup by Adobe Admins on the backend.
Salesforce Audience Studio
Salesforce supports an extensive list of partners for activation. It is also self-service where you can plug in account information and credentials but might require a support ticket to finalize it on the backend.
My POV
When it comes to activation, the key is the ability to confirm that the data is flowing properly with decent match rate. Both platforms offer these type of insights (Addressable Audience and Push Status) but Salesforce’s method is less confusing and offers more context that comes handy when testing and troubleshooting.
Testing and Troubleshooting
No one tells you how much testing and troubleshooting you have to do when standing up a DMP. It is often bundled under QA but for me troubleshooting is a phase that we must account for when setting up either platform. This is needed to confirm that tags firing on the site/app, segment logic and mapped audiences are setup and performing properly.
Testing and Troubleshooting
Adobe’s Visitor Profile Viewer is an excellent tool in AAM to troubleshoot trait and segment qualifications especially when testing site data collection. Along the same line, Adobe offers “Experience Cloud Debugger Extension” specifically for Chrome browsers and “AEP Assurance (formerly Project Griffon)” for mobile apps to help in monitoring and verifying data collection.
Salesforce Audience Studio
In addition to forcing a user into a segment in “test mode”, Salesforce offers a “control tag helper” javascript that can be bookmarked in any browser.
This tag is very helpful for validating and troubleshooting site data collection and tags in general.
My POV
I would like to see more tools from each platform that eases the process to validate, QA and troubleshoot without the need to open network analyzers and sniffing every call to check the request and response. A lot of effort goes into this phase of the setup and there is a lot of opportunities to make it better and easier. Maybe we can automate some API’s to regularly verify the implementation and troubleshoot any data loss?
There is a lot of detail in the setup phase and we barely touched on them all. I hope I was able to convey the important things to consider to give you the best overall understanding of their offerings. Next up is post-implementation!
