Seamless Journeys: Unlocking Identity with CJA

Transcript

Hi everyone.

We are just waiting for some more folks to join in. We’ll be starting this session soon. All right.

Hello everyone and welcome to today’s session.

We will be covering the identities. The session is on unlocking identities with customer journey analytics.

I’m going to go ahead and kick off the session for today. First and foremost, very heartful thank you for your time and attendance today. Just to note that this session is being recorded and the link to this recording will be sent out to everyone who has registered.

This is a live webinar. So it’s in a lesson only format, but do feel free to share any questions into the chat and the Q&A board. Nikunj and I will be answering them as possible. And in addition to that, we also have a reserved time to discuss questions. That may surface at the end of the session.

So my name is Mayank. I’m a multi-solution architect with Adobe and have a little more than 16 years of experience with Adobe products. Along with me is Nikunj and I’ll let him introduce himself. Hi everyone. Good morning, good evening. This is Nikunj. I bring in about 11 years of experience in MarTech solutions and about four years of experience with Adobe Stack. I primarily work as a senior technical consultant for RT, CDP and CJA.

Back to you Mayank. Thanks Nikunj. So in today’s session, we are going to talk about something that impacts every organization, understanding your customer across their entire journey. Whether you are from business or marketing team or analytics or engineering, this session is a bright redesign for you. Today, we will be simplifying what customer journey analytics really does, why identity stitching matters and how this unlocks a better decision across the teams.

While we wait, I want to let you know that we have several other sessions scheduled and delivered for this quarter and that all our prior sessions recordings are also available on the experience link and can be accessed on demand. Finally, we are just a few of our most recent sessions. If you are interested, please go through those. It will be quite intuitive and knowledgeable. Also, today’s session has been prepared for you by our subject master experts from Adobe’s Ultimate Success team. What you see on this slide is a foundation of our Ultimate Success model, which actually combines proactive strategy leadership with responsive technical support to help the customers maintain their stability with their Adobe solutions.

On the proactive side, we work hand in hand with your team to align on the unified success plan and support your roadmap through the expert-led activities. At the same time, we are here to ensure the rapid response when needed. Our responsive model includes dedicated support resources, subject matter experts who monitor, prioritize and resolve issues, whether it can be P1 incident or an ongoing incident analysis. Together, this approach ensures you are not only covered for today’s need, but continuously moving forward towards a long-term success and value realization. All right, so let’s talk about customer journey analytics. In simple words, CJ allows anyone, not just data analyst or data scientist, to answer complex business questions without writing SQLs. Traditionally, the data sits in multiple systems.

It sits in your online data sets, offline data sets. It can be through the email, mobile, web. And analyst writes complex queries and waits for days and weeks. With CJ, the data is connected with the connections and all the data based on the IDs associated with every data sources can be brought in together.

The insights and the visuals will be created on the data views, which are nothing but kind of dynamic data, which is available on top of the connections. And the exploration is drag and drop more intuitive. And for the technical team, it is empowered by Adobe Experience Platform. And for business, it feels more intuitive and interactive. It is quite powerful where the analysis can be utilized and activated to several other channels, which can again bring back the data into CJ for your further analysis and optimization. So now let’s discuss a problem with most of the organization’s face. Customer data lives in silos, website data, paid media, email, or a call center, any offline data, and they don’t naturally talk to each other. When your leadership asks, why did customer drop off before completing a payment or before completing a purchase? The answer isn’t always easy. CJ does solve this by creating sequential multi-channel views of the journey. And instead of isolated events, we see a clickbait search, browse product, account created, money transferred, and repeated transactions. Now we can identify where a customer falls out, which creates friction, which channel truly influence conversion. This means faster insight for the business users and technical users. This means a unified data model driving the same insight. And once we understand these journeys, the next big question is from the attribution. Attribution is always misunderstood. Most of the organization uses different models, different attribution models, in different systems. One model for email, another for paid media, and another one for the offline. And the results are always conflicting.

CJ changes this. You can apply attribution to any metric, just not to the campaigns, but to any metric which uses a look-back window, builds the models at the report time, and no complex implementation which is required. This again means marketing analytics and the rest of the team all sees the same data, same truth.

Again, attribution logic is flexible and applied dynamically, which helps our business team to get the consistent and trusted insight. But to make attribution truly accurate, we must answer one big question. Who is the customer? Who is the same customer across the devices and the channels? And that’s exactly where our identity stitching comes into the picture.

Let’s start with the simple one first, field-based stitching. This connects the data using a common identifier. For example, your same data or same data set or same data source uses a device ID or a cookie ID along with a login ID or a CRM ID.

If the same ID exists in this data set, we connect them. This allows us, basically this type of stitching allows us to restate the anonymous history once the user logs in. It also connects the web behavior with CRM data and it works well if the same ID exists everywhere.

But the real challenge in real life, customer uses multiple devices. There are multiple channels. They can always go back and try switching browsers, interact offline. So the same person may have multiple identifiers, multiple device IDs and that’s exactly where our graph-based stitching comes into the picture.

So instead, unlike field-based stitching, relying on only one shared ID, it connects identities using relationship. Think of this like device A is linked with an email X and email X again linked with a CRM ID Y.

So then the system understands that they represent the same person. It is like connecting a puzzle piece automatically, but technically it leverages the identity and the graph relationship which sits inside AAB.

And the advantages are significant. No need for same person ID across all the data sets, higher identification rates, alignments across real-time CDP and other app services like KHIO and more accurate data set joins based on the common identifier or the relatelet identifier. Again, the benefits can be better journeys, better attribution, better personalization and ultimately better business outcome.

All in all, to summarize, the difference is simply the field-based stitching requires a consistent ID and works with an additional ID, works when the data is clean and aligned. Graph-based stitching on the other hand works even when the IDs differ. It connects more data sources. It improves the accuracy. All in all, if your organization is scaling across the channel, multiple devices or the regions, graph-based stitching is the strategic approach. When we combine across channel flexible attribution identity stitching, we unlock the faster time to insight, increased efficiency, higher lifetime value. And this is not just analytics, but it’s a unified customer intelligence strategy. Having said that, we have covered the concept and let’s see how this looks in action. In the demo, we’ll walk through how data connections are built and how stitching makes an impact by connecting journeys. As you watch this, think about what business questions could I answer faster with this. Handing it over to Nikunj for the demo. Yeah, thanks, Mayank. Let me quickly share my screen. Yeah. Yeah. So, as Mayank already talked about the differences on field-based stitching and graph-based stitching and how we can leverage that, let’s get into the platform to sort of understand how we can create this and how we can utilize this within CJA. Right now, you see my CJA screen. I’m under the screen of data management and set connections. Let’s go ahead and create a new connection where we can leverage the identity stitching. So, here is where we would give our connection name. So, I’m going to say, demo webinar list and then you could give a description.

You would select the primary ID and then you can enable rolling windows based on how much data you want to get into CJA. Select which sandbox you want to get your data from and then also select the average number of daily events that is needed. I’m going to select less than 1 million. Now, here is where you start configuring your data sets. So, you click on this button on add data sets and you see a pop-up with list of all the data sets from the sandbox that you selected in your earlier configuration. So, I already have created two data sets on AP platform. One is an event data set and one is a profile data set. In my event data set, I have created two identities that I’m using. One is an EC ID and one is an email ID. The EC ID is what gets generated for anonymous browsing events and email ID is what you get along with the EC ID for authenticated events. So, all of that web data comes into this event data set and then I have a profile data set which has email ID and few attributes of that particular profile. So, let’s select our event data set here and we’ll click on next.

Now, here is where you see. When you don’t leverage any type of stitching, you would ideally directly select a person ID here.

In CJA Connections, when you’re working across multiple data sets, here is where you would select that common person identifier that was similar across all the different data sets. For example, if you are using an email ID that you are getting among your web data, transaction data, and CRM data, then you would go ahead and select that as a person ID that is common. In our case, we want to leverage the enable identity stitching. Now, this feature was generally available to all customers starting of Jan 30th. Prior to this feature, you would have to raise a support ticket with Adobe to get the stitching enabled. From Jan 30th onwards, this feature is now generally available where you can define the stitching on the connections configuration itself. So, let’s click on enable identity stitching. You will see a small pop-up here which talks about enabling stitching, what it derives, how is it used, right? So, you can just go through this message. If you want additional information, click on the experience leak documentation. And if you are well aware of the feature, just click on continue. And this is where you would get options to select, right? Now, let’s talk about the simpler approach of field-based stitching. In field-based stitching, this is as the name suggests, the stitching happens based on two different field attributes within your dataset, right? So, the way you would look to configure this is you would select a persistent ID. That persistent ID is something that would be available in most of the rows in your event dataset, right? Which is EC ID in our case because that’s the one that we are selecting the, that is available for anonymous users as well, right? So, EC ID is going to be that is something persistent. In your person ID is the ID that you want to select as the as the stitched ID which will be used and populated, be populated among all the other rows as well, right? So, in our case, we will select email address here since that is available. Now, in the same option, you can see an option of identity graph, right? When you select identity graph here, this is where it goes and starts using the graph-based stitching as Mayank was explaining earlier how it is leveraged, right? So, once you select an identity graph, you again see a pop-up window here that says ensure you have finished the setup of identity graph before you use graph-based stitching, right? So, in your experience platform, you need, you need to have the identity graph configured, all the identity namespaces configured in the right way for you to utilize that in CJA, right? For our demo, I am going to just click on email, use the field-based stitching, right? So, email is my authenticated person ID that I want to select. Then, you can select the replay window, right? You can select the replay window of how many days worth of data you want to go back and replay to use this for stitching. Again, graph-based stitching and the replay window, the options that you see down depends on the licensing that you have. So, based on the package of CJA that is there, you might not be able to see if you do not have either CJA prime or select, sorry, CJA prime or ultimate, right? For event datasets, it utilizes the timestamp field. Then, you select what data source type is this event dataset. I am going to select web data for my case and then you move on to import new data, right? You always want to check this box in case you want to refresh this event dataset. So, I am going to click on import all new data and then I am going to also enable dataset backfill, like how much of data that is already existing in your dataset you want to bring on to this connection. So, if you want all the existing data, you can click on this toggle saying backfill all existing data. The backfill in itself gets queued, right? So, in case you want to request a backfill for a specific timeframe, you can go ahead and select that here as well, right? And then you can queue that backfill that I want only last 60 days of data or whatever, right? So, once you have configured all of this, you click on add dataset and then you save your connection, right? If you have more datasets, you can add that as well and then you can see here within the connection, the stitched column will be set as true based on the since we have used field based stitching, right? And when you go into the dataset settings as well, you will be able to see the person ID is now utilizing a stitched ID, ok. So, for our demo purposes, I have already created the connections and I have kept, I have created a connection for field based stitching that we will leverage in our workspace reports to sort of understand how the stitching works from a data standpoint, ok. Now, let us move on to the field based stitching in the workspace and try to understand how these events actually stitch and what is how do we see the benefit by utilizing this feature, right? Now, here is one freeform table that I have created, right? This is a very simple report. When I have created a metric called has email, what this has email metric is? Anyone who has an email identity set, so this is a metric that I had added in the data view.

So, if you see here before stitching, right? For the month of January, I have 73 total events that are there out of which 58 event rows have email ID attached to them. Similarly, for February, there are 35 events in total and 26 have total number of email ID’s attached to the event rows, right? So, this is something that we see for without stitching. Now, how does stitching help us, right? Now, I have created certain calculated metrics here to help us understand more details, right, of how this is getting stitched, right? I have just expanded the earlier freeform table below here. I have added few more columns here. The first calculated metric that I want to talk about is email stitch, right? Now, what this calculated metric is? Let me just put it right here. The wrong one. Yeah. So, in this calculated metric, what I am doing is I am calculating every all data that has an identity namespace equal to email and then using event here, right? Event container within this.

So, all I am counting all the events that have an identity namespace of email in it, right, after the stitching is used. So, if you see here earlier out of 73 events, we had 58 events that had email, but with the stitching what we were able to do is we were able to add 10 more event rows using the failed based stitching and now 68 of our event rows have email ID to it, right? So, this is the uplift that we are seeing based on the failed based stitching that is utilized. Similarly, for in the month of February, we had 26 event rows out of 35 that had email address earlier and with the stitching, with the failed based stitching, we see that the count has increased from 26 to 35. There was no change in November and in December we see 7 more rows had the email ID attached to it through failed based stitching, right? This is the uplift that we saw for to get more understanding on how this looks from a percentage perspective, I have created two more calculated metrics. One is the authenticated standard authentication rate and email authentication rate. In the standard authentication rate, it is a very simple calculated metrics. Anyone who has an email divided by total number of events make this as percentage at two decimal places and then similarly for the email authentication rate, stitch authentication rate, we are using all the events that now have an identity namespace of email divided by total number of events. So, these are the two calculated metrics that we are going to use. These are percent metrics. This will just help us understand in terms of percentage, what is the lift that we saw with total number of events that we have versus how much increase we have seen, right? So, if you see here for example, before stitching, before the identity stitching, 79 percent of our events that we got in the month of January had an email address attached to it. With the identity stitching, field based stitching, we were able to increase that number to 93 percent. So, after the field based stitching, we see in the month of January, now 93 percent of our event rows have email ID attached to it, right? Similarly, for February, we see 74 to 100 percent. All of our events were able to be 35 events we had in total and all 35 were now having the email. So, we see a 100 percent identity stitching happening here across all events for the month of February. We do not see any change in November because there was no events that were able to be stitched and for December, we see an increase from 64 percent to 87 percent. So, these are some of the quick validation reports that you can utilize in your connection to sort of understand how field based stitching is working and versus how it is being leveraged and what is the uplift that you see once you have enabled this feature, right? You can do this for your event data set for your web data sets as well. You can create this simple calculated matrix to get an understanding of the total number of events and what is the lift that we have seen. Yeah, that is it from the demo perspective on the stitching side. Back to you, Mayank.

All right, great, great. Thanks Nikunj. So, just to wrap up this CJ, what helps you with is to break down the silos, understand the identities. You can align the business and technical teams, make it faster and more smarter decisions. As you have seen how it increases the identified number of rows or the events within the CJ and with the replay in place, it will increase more further.

Whether you are marketer, analyst, architect, executive, the goal is always to see the customer clearly and act confidently and deliver the better experience. And that you can do once you identify the customer across the channels, across the devices.

So, having said that, we like to thank you for your time today. We are getting into a Q&A portion of our event. There will be a small quick two-question poll launching to get your feedback and help shape our future sessions.

Please feel free to drop your questions if you have any in the Q&A section or in the chat box. Thank you. All right. I think we can close the session and thanks once again for your time today. Have a great day. Thank you. Thanks everyone. Thank you.