Adobe Audience Manager (AAM) and Salesforce Audience Studio (Krux): Side by Side Comparison: Part I – Data Collection
I started writing this series back in September 2018 but didn’t feel the urge to release it until I saw Forrester’s latest DMP wave where it identified the DMP offerings by Adobe and Salesforce to be the only leaders in the industry from strategy and offering point of view. I respect the work that Forrester and Gartner put into releasing these type of reports. There is a lot of research, testimonials and work that goes into this scoring system. That being said, I am never convinced that they paint the full picture for the practitioners and the folks who work with these tools once the contracts are signed and the sale is complete. Therefore, I decided to take my experience in both platforms and put it into a four-part series:
First things first, if you are new to Data Management Platforms (DMP), it is OK. I recommend watching this 2 minute video that summarizes what DMPs are. If you ask me, a DMP has three main components:
- Data Collection: a consolidated platform that combines 1st, 2nd and 3rd party data together. The data hub that brings online and offline data under one roof
- Audience Aggregation & Segmentation: Segmentation is a key feature of DMPs where segments can be based on data attributes from various sources (online and offline)
- Audience Activation: DMPs are integrated to numerous Adtech and Martech vendors where these audiences can be accessed. These integrations vary and range from browser/server side to real-time/batch process and allow marketers to target their audience segments on multiple fronts
Second thing and before we jump into the Implementation details. It is important to note that Softcrylic is a technology agnostic consulting company. We partner with every technology, including Adobe and Salesforce, to stay up to date on their offerings and have the right level of expertise to help clients throughout their journey from tool evaluation, implementation, training and post-implementation strategy. If you checked my bio, you would know that I spent 6 years of my career at Adobe before joining Softcrylic. A good chunk of this time at Adobe was on Audience Manager (AAM). I was fortunate to work on really complex implementations that were not your common “plug and play”. This set me up to work with product, engineering and consultants. Before I knew it, I was the liaison between them and we were coming up with solutions that would get productized quarters later. This allowed me to learn the ugly details of data ingestion, edge qualifications and the data architecture landscape that a DMP was built on. Ok, so you know AAM well but how about Krux? The simple reality is that the DMP technology is not complex. Not to undermine the power and value of a DMP but deep down all DMP’s are similar: Collect, Segment, Activate. After working with Krux for the past 2 years, I am comfortable in formulating my POV on both products.
Enough with the introductions and history, now let’s get down to the details.
In the implementation phase, I want to concentrate on the following (split across 2 parts):
Real-time Data Collection
Needless to say, the biggest selling point of a DMP is the idea of collecting behavioral data (Web, App and Mobile) in real-time. Both Adobe and Salesforce do this well but a bit differently. Let’s dive into each:
Adobe Audience Manager
There are two ways you can send data to AAM from a web property. An old but still common way requires you to deploy DIL (Data Integration Library) on your site. DIL is a javascript (JS) library with a handful of functions that allow you to instantiate ID Syncs, hook into data objects and send data in a defined key-value pair format. Now if you also happen to have Adobe Analytics in your stack then you can enable server-side forwarding (SSF) that eliminates the need for DIL and instead enable an AAM module within Analytics and voila, all your analytics data flows instantaneously to AAM.
In terms of the mobile app, the most common and supported way would be to deploy the Adobe SDK. This also has two options, Adobe SDK v4 or the new AEP v5 which allows you to use Adobe’s Tag Manager (Tags (formerly Adobe Launch)) to hook into the app and make some updates without the need for a new build. Needless to say, I have seen other clients use Tealium and Ensighten to implement the Adobe SDK in their mobile app. It really depends on the app design and the development cycles.
Salesforce Audience Studio
Today Krux offers one way to collect data from your web property. This is enabled through a control tag. The control tag is also a JS tag that instantiate the library, parameters and data transfer code (DTC). Many things are configured within DTC but the important one to highlight is the data layer. Krux collects out of box variables but also allows you to hook into a data layer that can bring in data automatically.
In terms of the mobile app, Krux has an SDK for iOS and Android that requires tagging the app to track views and events. It is also common to see Krux implemented as a pixel within a tag manager on a mobile app if the size of the SDK is a concern.
My POV
As you can see, both tools are very similar when it comes to data collection from web/app. You still need to use a tag manager, open a ticket with client care/service desk, configure JS code and a server call is fired from the browser. That being said, there are two differentiators that are worth pointing out:
-
Ease of Deployment
Adobe’s integration between AAM and Analytics is undeniably a strong contender when it comes to data collection. You don’t need to reinvent the wheel in tagging the site if it has already been tagged with Adobe Analytics. On the other hand, Krux might require more development work to setup the data layer and setup the pixels needed to capture events off the site. In situations where a client utilizes Google Analytics, there are ways to expedite Krux’s deployment by hooking into Google’s dataLayer and leveraging GTM’s events to trigger Krux events.
-
Media and Email Tags
Both DMPs support these types of tags (pixels). It is a 1×1 pixel either trafficked in the creative of a display tag and fires for impressions/clicks or embedded in the email and fires on email open/click (for browser based emails that supports images). For AAM, you can find documentation on the microsite that outlines the structure of these pixels or ask the consultant to create them. In Krux, you can create these tags directly in the UI and there are templates for all the common ad servers which makes the process super easy.
ID Syncs
Also known as User Matching is the mechanism of synching the DMP ID with another ID (DSP, Data Provider or Hashed Customer ID). For Audience Manager, Vendor ID Syncs are enabled on the backend by an Adobe Admin or Adobe Support and triggered via an iframe while 1st party ID Syncs are configured in the tag manager and require data source or integration code setup. In Salesforce’s case, vendor ID Syncs are enabled by the service desk while 1st party ID Syncs are configured within the control tag and require data layer setup.
My POV
Although both platforms support the ability to have more than one ID Sync enabled for first party, there are few key things that are slightly different:
-
Frequency for 3rd Party IDs
In AAM, it is guaranteed that ID Syncs fire on the first pageview of the first visit as well as once every 14 days. In Krux, user match pixels are fired a max of three times per week per device but it is not guaranteed that they will fire on the first pageview. There is a priority tier set for different vendors and based on that some tags fire on the first pageview while others wait for the second page.
-
Reporting for 1st Party IDs
Krux offers a report (First Party User Match) which has a 30 day and 90 day lookback count for every first party ID that is enabled for a user match sync. In the last year, Adobe released new set of traits that are automatically generated based on data sources which basically does the same thing. It is not exactly a report but by trending these traits you will be able to see the number of unique IDs synched on a given day.
Data Onboarding
Maybe the biggest selling point for a DMP: “Onboard your offline first party data and stitch it with online behavioral data”. It is also the biggest hurdle to accomplish. Both DMPs follow the same procedure:
Both platforms need an sftp or s3 location setup for the data to be dropped into prior to ingesting it once every 12 hours.
My POV
Adobe’s onboarding process is much smoother and for 3 reasons:
- Onboarding status: Adobe offers a report that shows you the historical status of onboarded data over time. It even allows you to see a sample of the bad records which is essential for troubleshooting.
- Email Receipts: You are onboarding a customer file right before a big campaign and you are not infront of the platform. How about getting an email notification outlining how many records were ingested and how many were rejected? Trust me, it is a big deal.
- Overwrites: AAM allows you to upload incremental and full refresh files. This is defined in the file extension (.sync vs .overwrite). On the other hand, Krux only supports full refresh. This could be a big deal if you are uploading a large dataset and only few values are changing over time.
This concludes the first post in this series where we covered how each platform ingests data in realtime and batches (I will leave API to the end ☺) Unfortunately, neither platforms support a drag and drop mechanism for data collection or uploads but maybe that is something we see in the future. That would be nice to drop a file and see data instantaneously no?
