Vídeo em Inglês, publicado pelo canal Lukas Bihary.
About this content: The summary, key points and topics below are generated automatically from the video transcript. They reflect the content, opinions, claims and predictions presented in the video and do not represent a position, validation or fact-check by All The Gate.
Video summary
The video discusses the essential technology stack required for personalized e-commerce experiences, focusing on companies with annual revenues between 25 million and 2 billion. It breaks down the tech stack into three main interconnected regions: data tracking infrastructure, data activation infrastructure, and data storage infrastructure, explaining the function and importance of each.
Key points
- The e-commerce tech stack is divided into three regions: data tracking infrastructure, data storage infrastructure, and data activation infrastructure.
- Data tracking is the backbone for analytics and activation, and digital identity management is a key challenge due to cookies and multiple devices.
- First-party scripts help extend cookie lifetimes compared to third-party scripts, minimizing fragmented customer profiles.
- Data activation infrastructure is often centered around a tool like Bloomreach Marketing, which includes a fast CDP for real-time personalization and multi-channel marketing.
- Data storage infrastructure, such as a data warehouse (e.g., BigQuery), is necessary for cost-effective long-term data storage and non-customer data like ad spend and inventory.
Topics
e-commerce tech stackdata trackingdigital identitycustomer data platformwebsite personalizationdata storagemarketing automation
What you will learn
- E-commerce tech stacks for mid-to-large companies can be structured into data tracking, data storage, and data activation regions.
- Visitor behavior can be split across multiple profiles if users switch browsers or devices, which can be mitigated by using common identifiers like email addresses.
- First-party tracking scripts generally allow cookies to last over a year in browsers like Safari, whereas third-party scripts often expire in seven days.
- A fast database like IMF inside a CDP is crucial for real-time website personalization, while a data warehouse is used for cost-effective long-term storage and broader analytics.
Questions and answers
What is the recommended annual revenue size for the companies targeted by these tech stack recommendations?
The recommendations are suited for e-commerce companies generating between 25 million and 2 billion in annual revenue.
Why is digital identity split into multiple profiles in a CDP?
Digital identity is often split because cookies are assigned to specific browsers on specific devices, meaning a user visiting from multiple devices or browsers gets assigned different cookies and profiles.
Why is a separate data storage infrastructure needed if a CDP like Bloomreach Marketing is already used?
While CDP databases are very fast and necessary for real-time personalization, they are expensive; therefore, a separate data warehouse is needed for cost-effective long-term data storage and non-customer data like ad spend and inventory.
Video transcript
Transcript from available captions; it may contain errors.
So, in the last couple of videos, we talked about how a personalized shopping experience can look like in the e-commerce space. And today, I would like to talk about a technology that is needed or at least recommended to create such experiences. I would also say it heavily depends what [music] is the e-commerce size and the technology recommendations would be different. Uh for our specific case, we are helping e-commerce companies that do generate somewhere in between 25 [music] million to 2 billion in annual revenue. So, the recommendations that we will be talking about today are suited to the size of the companies. So, in general, I would say that we don't necessarily think that there is like a central point to the tech stack itself. We like to think about it as three different regions that fulfill a different role yeah are very important. These three regions basically communicate with each other. And here we prepared a nice interactive framework where you can see all of these three regions which are data tracking infrastructure, then data storage infrastructure [music] and data activation infrastructure. Okay, so let me quickly explain what each of these infrastructures are and what is their main function. All right, so let's start with data tracking infrastructure region first. The reason this region is so important is that data is really a backbone for everything that we are doing later on, either when it comes to data analytics or data activation. And yeah, like if you simply do not collect all data that is available and there are real gaps in the data that you are collecting, those gaps will be apparent in the data analytics and the data activation as well. I would say that there are common problems issues that needs to be addressed when it comes to data collection in e-commerce. And I would like to just explain to you from a high level what these problems are and then what you can do to address them. Cool. Before I do that, there is one really important thing that needs to be understood about e-commerce, which is digital identity. And I will explain it to you by again visiting our demo store datacop.com, which I did from the incognito window. And if I click on this button, you can see that I was assigned a visitor ID. [music] This is a lot of times referred to as a cookie as well. And if I copy this cookie and if I go to our CDP, which in our case we are using Bloomreach marketing, you can see that I can find a profile under this cookie. The thing to realize here is that this profile does not automatically represents me as a person because cookie is assigned to a specific browser on a specific device that is visiting your store. And you probably start to see a problem here because if I visit the store from a different device or even from a different browser, I get assigned a different cookie, meaning my website behavior is basically being scattered across multiple profiles. So there is no like a single profile that is representing my website behavior or my preferences as a customer. Typically, there are multiple profiles and my behavior and preferences are split across them. Of course, there are strategies that you can employ in order to minimize this split behavior across multiple profiles. And I will go deeper into this what these strategies are in later videos. But from a high level I will give you one uh which is basically customer need us to provide a common identifier across all of these profiles and therefore we can merge them based on that identifier. [music] And this identifier is typically an email address. So it makes sense to give a uh website visitor who we don't recognize under an email address to give them a strong enough incentive to tell us who they are by providing their email address. And we have a couple of use cases live on datacop.shop uh in order to address this specific issue. So for example, here we have a member exclusive products that can be seen only by people who we know under their email address. And same goes here for member exclusive discounts. And this is the primary reason why we have it there. So this digital identity uh issue is minimized to greatest extent possible. Another important thing to realize here is that cookie has a lifetime and this lifetime is decided by your browser, meaning Safari decides for how long this identity is stored in your browser. And once it's gone, even when you visit the website through the same browser, the website sees you as somebody who is visiting the website for the first time. Therefore, it also creates a new profile in a CDP. Uh now how Safari decides uh for how long it will store that cookie. And the number one condition is that is your tracking script uh categorized as a third party or as a first party? Third party script is basically a script of another company that is simply installed on your website, something like Facebook or Google tracking script. And first party should be genuinely your script running on your infrastructure. And the difference is huge. If you are running a script uh that is uh genuinely first party, Safari expires this cookie uh after more than a year, but if it's a third party script responsible for tracking, it is being expired by seven days. Meaning if you come back after eight days on the website via the same browser, it will create a new profile in a CDP and your identity will be split. All right? So the next region I would like to focus on is the data activation infrastructure. And typically in the setups that we are recommending, this region is dominated by a single tool that is called Bloomreach Marketing. And the reason for that is that from our perspective, Bloomreach Marketing basically has all the tools that are needed for the data activation across all channels that are relevant to e-commerce. What do I mean by this? So in a center of Bloomreach Marketing, there is a CDP or a customer data platform. And what is really unique about this CDP is that it is built on a custom database called IMF and this database is incredibly fast, which is very important for things like website personalization because if I visit the website, I basically need to ask a CDP like, hey, here is a cookie, who is this customer? What do you know about this customer? And then also decide what type of experience do I want to show this specific customer? And that needs to happen really, really fast. And that is typically something that your data warehouse cannot really support at that kind of speed. That is why it is super important to have something like this if, for example, website personalization is something that is important for you. What is really cool about Bloomreach Marketing is that it also contains all these other channels. So you can think of it like you also have a Klaviyo in it. So you can send emails through it, you can send SMSs, you can send mobile push notifications, you can personalize the mobile application itself. You can be sending all these is to pay channels such as Facebook, Google, etc. So from our perspective it is really a complete package that is tightly integrated together. Sure like you can go and you can buy different point solutions for every of these channel, but the downside is that now you have different systems. Every of that system has a different data structure and you also need to maintain the connections between these systems. And from our perspective it is simply easier like not to you know take care of these problems because Blue Ridge Marketing is solving these problems [music] for us. Okay but now you might be asking if I already have a CDP in a tool such as Blue Ridge Marketing, why would I need a separate data storage infrastructure? And the reason is very simple. As I already mentioned the database in Blue Ridge Marketing is quite special because it is very fast, but it comes with a cost and I mean that literally. Like these type of databases are absolutely needed for use cases such as website personalization. On the other hand they are way more pricier compared to let's say a data warehouse in Google Cloud Platform. A good analogy here can be your computer for example. If you think about it your computer also have two type of storage. It has RAM and it has the SSD. And RAM is responsible for all the programs that are currently running on your computer and on your SSD you have things like photos or films that are stored there for the time that you actually want to watch them. And this is exactly the same like for the data activation it is good to have that speed because it is needed for certain use cases, but it doesn't mean you have to have store all the data there because it would be just ridiculously expensive. The good alternative here is that okay keep everything that is needed for the data activation in a tool such as Bloomreach Marketing and keep everything rest in the data storage infrastructure in form of a data warehouse such as BigQuery. As an example, like if you have like a browsing behavior two to three years ago, it just doesn't make sense to have it in the data activation layer, but it is still very much valuable for the analytics purposes [music] if you want to do like year-over-year comparison somewhere else. And a good place for it is this data storage infrastructure. All right, and one more reason why you would like to have this data storage infrastructure is that CDPs are focused on customer related data, but in e-commerce there are more types of data. So as an example, uh you probably want to have access to ad spend across multiple platforms such as Google, Facebook, etc. You probably want to have access to things like your inventory, stuff like that. And you can centralize this in this one place which is the data storage infrastructure, which is really good for analytics. And in today's day and age, you can also do things like connecting your cloud directly to a data warehouse and you can start asking analytical questions directly against your data that is centralized at one place which is really neat. Cool. Yeah, so this is it when it comes to the technical marketing tech stack and how we are typically thinking about this. Uh this framework is actually part of Datacorp Academy that can be found on our website when we go even a little bit deeper into these topics. It is absolutely free of charge for some of the lessons you need to request the access, but we don't charge anything for it. All right, that's been it for this video. I hope you find it useful and see you on the next one. Bye.
Assistir ao vídeo original no YouTube
Assuntos: E-commerce, Marketing e SEO, IA e tendências. Publicado no YouTube em 2026-11-17.
Seleção automática por assunto e idioma. O vídeo pertence ao canal de origem.
Comentários
Postar um comentário