Have you ever received a request to track and detect online shopping cart abandonments in real time? If you have, then you are not alone. This is a typical request we get from our clients and I have seen too many times. The theory is very simple: if we can detect that a user has added something to the basket but has not purchased it, then we need to persuade him to finish the process. However, the reality is more complicated than just that. Let me explain what I usually discuss with my customers and what options do we have.
According to some news, all major Internet players are now focusing on chatbots. I have never used one but it looks like with the progress in artificial intelligence, we will all be using them in the not-so-distant future. Facebook even claims that thousands of developers are creating right now chatbots. If this is true, then we should be ready for them.
While I read this news, I thought, how about using Adobe Analytics to track the conversations? Would it make sense to large corporations, which already have Adobe Analytics, to use the same tool as with website and apps? I know some people will contend that my idea is wrong, that chatbots will need a different reporting tool. However, I would then reply that, a few years ago, it was not clear whether Adobe Analytics would be used for apps; now, all my clients, want to integrate Adobe’s SDK in all their apps.
If you are working with a DMP like Adobe Audience Manager, I am sure you have come across the following problem: you want to target your visitors on site, immediately after they log in, using on-boarded data, even on the first visit. This last statement is, precisely, where the problem is. The way AAM processes on-boarded data is as follows:
- You upload your CRM data to AAM, either to an SFTP location or an S3 bucket
- Every 12h, AAM reads all on-boarded data and processes it, converting the signals into traits
- The traits are stored in the core servers
- A visitor logs in for the first time
- Since the communication between the browser and AAM is done through the edge servers, these servers have at this moment in time no on-boarded information for that visitor
- The edge servers where this visitor activity has happened, request the on-boarded traits to the core servers
- In a batch process, core servers send to the edge server the visitor’s on-boarded information
Initially, when we think of segments (or clusters) for ad segmentation, we think of ever-growing groups of cookies. Simple use case like purchasers, visitors to our website, subscribers to a newsletter or owners of a device fit in this model. However, advanced (and not so advanced) use cases do not work well with this model, where we have visitors entering and leaving regularly a segment, so a segment can shrink in size:
- Retargeting dropped baskets: the moment someone places an order, you do not want to retarget him again
- Customers of a mobile operator: it is very common nowadays to switch to a different provider frequently
- Age group: ever day, visitors enter one particular age group or leave it, as people grow older
My job as an Adobe Analytics consultant has involved, very often, bridging the gap between these two worlds. I have seen myself many times as a translator: getting a message from the marketer, translate it into technical terms and communicating it to the developers; and vice-versa. My developer background has helped me a lot in this case. In quite a few cases, I have been requested to join meetings just to make sure that the IT team understood what the marketer wanted. It does not help either the fact that web analytics is not considered as important as it should be.
Let me start with an anecdote I once heard. The marketing department decided that they wanted a new feature in the home page. The IT team received the request and implemented it as per the requirements. Three months later, the business owner of this new feature requested a report on the performance of this new feature to the web analytics team. To the team’s surprise, that was the first time the web analytics team had heard of this feature. Consequently, had not issued any tracking requirements and there was nothing to report on. In other words, three months had been lost.
Before I started working with Adobe Audience Manager, I had a very limited knowledge of the on-line advertising market. In the past, I had managed Google AdWords campaigns, but that was all I knew. Now that I have been working for some time with a few AAM customers, I have realised that the market for on-line campaigns is huge. There are many actors involved: agencies, trading desks, DMPs, DSPs, SSPs… I still have to learn more about this market.
Today’s post is going to be a different form the last few posts, a bit more hands-on.
One of the typical questions I get from my AAM customers is “how do I detect a user browsing with an iPhone [model]”. The only solution we have to reliably detect the device is through the User-Agent. Although this should be very simple, in theory, there is one problem: Apple does not want you to detect the iPhone model. Android devices include in the User-Agent the name of the device, or enough information to get it from there. However, Safari browsers include the device type (iPod, iPad or iPhone) and the iOS version, with no hint of the model.
This is going to be a rather short post, but only from my side, as the poster: if you follow everything I am saying, it will be even longer for you to process than any of my previous posts.
Let’s start by watching the one of the great TED talks: Sebastian Wernicke: How to use data to make a hit TV show
This is my first attempt to write an opinion article. I had it in my mind for some time, but the sparkle was a question during my talk at the London Analytics Labs. One attendee asked me about the future of on-line advertising if 3rd party cookies and/or ads were blocked from all browsers. So, this is my point of view.