Twilio Segment

My main use case for Segment is to share a practical experience about how we used Segment at XP Investimentos bank to transform behavioral data into business intelligence and high-impact marketing campaigns, where we worked in the MarTech arm with a very clear goal: to turn Segment into our CDP, making it the main source of truth about our customers while understanding how they interacted with the XP app to personalize the end-to-end experience.

Before giving a specific example of how I used Segment to personalize customers' experience in the XP app, I will talk about the challenge and the technological solution, where my role as a software engineer was directly tied to implementing, maintaining, and optimizing this pipeline. I knew the tool very well inside and we created and configured sources, structured the receipt of events via HTTP calls or webhooks, and developed specific scripts to trigger two fundamental types of events: track, focused on capturing behavioral events within the app, and identify, responsible for capturing customers' registration data, ensuring we continuously used and updated customers' registration data for a solid, enriched, and always up-to-date base. The architecture of the solution and the data flow was the way to make this work at bank scale, managing millions of daily events through a robust architecture where events generated in the app passed through an intermediary application responsible for collecting these actions through messaging services like Kafka and RabbitMQ, allowing us to format the events with custom fields for proper identification, processing, and standardized sending to Segment, which let us segment and create highly qualified audiences crossing in-app browsing behavior and product acquisition with the client's investor profile, opening a huge range of operational and marketing opportunities like abandoned cart campaigns and personalized banners displaying real-time customized offers.

The process to create those qualified audiences involves a multidisciplinary approach seeking a successful final impact, where building this ecosystem is not a one-person job and requires deep integration between multiple technology and business fronts, as we worked side by side with the Mobile, interface, hub, advisors, BFF, backend, and database teams to ensure that front-end collection was performant and did not impact the app experience, while serving the Revenue team directly as it used Segment as its main work tool and source of truth, enabling them to create dynamic audiences, trigger targeted campaigns, and extract precise reports for strategic decision-making in a summarized implementation of Segment at the XP bank, transforming our communication approach into a fully personalized strategy focused on the investor's journey, an enriching technical project where I applied software engineering to leverage MarTech results generating real business value.

I believe that the offer banners were successful cases from my experience with Segment, as we tracked the entire flow from click, view, and conversion of marketing campaigns, which brought in significant revenue reaching about 650 million in monthly revenue.

In my opinion, the best feature that Segment offers is the creation of audiences focused on the investor's profile, which was the biggest differentiator enabling us to cross a large number of behavioral and customer profile data points, generating increased engagement in marketing campaigns.

Regarding audience creation, I consider Segment CDP interface itself the most valuable feature, as it allows the marketing team to work with a business-focused profile without needing deep programming knowledge, greatly facilitating the creation of marketing campaigns while integrating with tech solutions.

Segment allowed for the organization of the entire source of all audiences, making it a very robust tool where the cost pays for itself through return on investment, highlighting the tool's cost-benefit ratio as a major differentiator.

Segment has a very significant positive impact on my organization as it provides robust data to understand our consumer's profile and needs, focusing on the customer's needs in various life stages, ultimately maximizing our efforts on the customer, which is where the main gain with Segment lies.

I have not used Segment for more than a year, so I am not aware of its current evolution, but I know there was a migration from Segment to Adobe as a strategic decision by XP bank, which pointed out unified databases as an improvement area compared to Segment, where they effectively utilized AI for creative marketing campaign generation.

The necessary improvements in Segment based on my experience focus on providing more tools for the marketing team, particularly using AI to create campaigns through images and short videos, enhancing communication with end customers via offer banners in a more automated and effective manner.

I used Segment from December 2022 until around the end of 2025.

Segment is very stable with no major issues encountered, aside from avoiding Friday deployments to prevent losing any event capture, ensuring stability throughout our operations.

Segment's scalability is very good given the large processing capacity handling millions of events from around 10 million customers daily, where every in-app navigation and various accesses were tracked and sent to Segment CDP without encountering issues.

Customer support for Segment was satisfactory, as communication happened through tickets and scheduled meetings with support staff who sometimes spoke Portuguese, with a need to adapt to the tool's capabilities rather than having fully customized solutions. I would rate customer support around seven or eight due to what I perceive as limited availability for immediate assistance compared to the overall value derived from the tool's capabilities.

I did not use a different solution before joining XP as we directly employed Segment CDP, which was later replaced by Adobe.

My experience with pricing, setup costs, and licensing was smooth, though I did not know the total expense, estimated around 1 to 2 million per month, with licensing restrictions on access to Segment CDP tool, experiencing a negative aspect when triggering numerous tests that exceeded our monthly limit resulting in a 5 million fine.

I have already shared measurable results demonstrating this positive impact of Segment at XP Investimentos, highlighting our ability to increase engagement and conversion in marketing campaigns by 30-40% depending on the campaign, generating monthly revenue of up to 650 million through triggers sent via email, WhatsApp, and push notifications in the XP app itself.

I have previously detailed the return on investment from our relevant increase in revenue attributed to Segment, where for every 1 to 3 million invested, we generated 650 million, with no staff reduction but an increase of advisors from 17,000 with goals now surpassing 20,000, while we also aim for a target of 2 trillion in investments under custody, highlighting our solid knowledge base for relevant marketing campaigns.

My experience with pricing, setup costs, and licensing was smooth, though I did not know the total expense, estimated around 1 to 2 million per month, with licensing restrictions on access to Segment CDP tool, experiencing a negative aspect when triggering numerous tests that exceeded our monthly limit resulting in a 5 million fine.

I do not recall which other options were evaluated prior to Segment, but I recognize that Adobe was in the same tier, and the challenge was that multiple applications had already integrated with Segment, leading to a migration project lasting about six months.

My advice for those considering using Segment is to thoroughly research the market and understand the tool's benefits as it is a strategic investment that necessitates discussions with other businesses to benchmark their experience and success post-implementation. My final comment on Segment is that it was a very enriching experience that taught me about event-oriented applications and how behavioral events and registration data are stored in Segment, allowing me to appreciate the potential for creating marketing campaigns based on a deep knowledge of our customer profiles. I would rate this product an 8 out of 10.

My main use case for Segment is to perform ETL with tracking user behavior, user funnel interactions with the website, where the user dropped, and bringing all this data into our database.

A specific example of how I use Segment for tracking user behavior or funnel interactions involves using a SQL database, particularly a Postgres database, where I ingest user tracking behavior that appears on the platform, such as what pages the user visited, what pages they dropped from, how far they dug into the website, and the items they clicked, as well as where they lacked interactions. This helps me understand how to reach users to make them complete the user journey, ultimately trying to encourage them to purchase our product. I collect all the data in our database to provide holistic insights to stakeholders and higher management about user behavior drops, guiding further analysis or actions to bring users back to complete their experience.

Regarding my main use case with Segment, we utilize GTM and HTML tagging to track users from multiple different channels and platforms, allowing us to identify the first point of interaction or touch channel details. This helps us know which platform or channel attributed to bringing the user to our site. We also create unique user IDs to analyze if one user comes from multiple different channels, allowing us to use either the last or first interaction according to our business needs. We then analyze this data through SQL or visualize it with tools such as Power BI or Tableau, which are powerful for gaining multiple different insights, demonstrating the overall BI use case of Segment.

The best features Segment offers include holistic interaction with different channels and platforms and the ability to bring data wherever I want, such as to any relational database. This is particularly useful because Segment supports analytics for GTM tagging and BI use cases, allowing user interaction data to reach our database.

Of the features related to integrating with different channels, flexibility in sending data to various databases, and supporting analytics and BI, I find that the overall combination is what makes Segment most valuable for my team. If I were to focus on one specific aspect of Segment, it may be comparable to other third-party tools available in the market, but what differentiates Segment is its integration of analytics and business intelligence into one platform, enabling us to derive different insights and understand what went wrong or right in our processes.

Segment has positively impacted my organization because we are getting the data correctly, and the license fee is also very much comparable with other tools, resulting in an overall win-win situation.

To share more about the positive impact, since integrating the BI part with Segment, it is really helpful for understanding the user journey and making useful decisions based on that journey. We can now focus on generating more revenue by identifying the top five or ten regions, channels, or platforms, allowing further bifurcation of user demographics through different analyses via SQL. Segment serves as our source of truth in our database.

Segment can be improved in the initial phase regarding data quality checks between different channels' attribution versus Segment itself. I have observed many cases where the data is incorrect or duplicated, often due to multiple channels attributing to one single user. Bringing the data into a unified understanding requires business-specific insights that are not particularly addressed by Segment. Although data accuracy was initially lacking, we managed to clean the data after it was brought into our database, which is a downstream responsibility rather than Segment's.

To elaborate on the needed improvements, data accuracy needs attention because there are use cases where the data is duplicated and requires deduplication. Additionally, the attributions or conversions from Segment also need to be accurate.

I have not been using Segment for many years; it has been very recent, with my experience being approximately eight months to one year.

Segment is stable.

Segment's scalability depends on the licensing, as more users or higher website traffic can be integrated into Segment with a higher licensing tier, which incurs additional costs. Nevertheless, the available upgrades and features in Segment are quite good.

I have not used customer support during my time with Segment, so I hope it is reliable. However, I am uncertain if customer support is now handled by AI or real-time human agents.

I previously used an in-house solution to bring data into our database and did not use any third-party tools. This marked my first experience with a third-party tool, Segment, which we opted for after conducting thorough research that indicated it would be our best current solution.

I have seen a return on investment after integrating Segment, with significantly increased revenue due to the introduction of BI capabilities, where we pull data from the website to our database. Revenue serves as a direct metric, while the BI part further enhances revenue outcomes. Additionally, our developers use Segment more consciously and precisely, and with the help of AI, we anticipate reducing costs further in the future.

My experience with pricing, setup cost, and licensing is primarily based on comparisons with other tools, revealing that Segment is much cheaper, particularly regarding premium licensing.

Regarding Segment's AI capabilities, I believe governance and security involve a worldwide unified approach being monitored across different tools and platforms integrated with enterprise data. AI governance and security need to be assessed properly before AI can take over various functions on our behalf.

On the topic of AI accuracy and reliability of output, I have not extensively used AI since it was not integrated much when I left my old organization, but I hope AI can significantly enhance the platform's capabilities compared to those without AI.

My advice for others considering Segment is that if you want a unified solution for analytics and BI, Segment should be your go-to choice.

I rate this product seven out of ten.

Segment connects with our data streams to send WhatsApp campaigns and email campaigns to our international audience, including Argentina, Nigeria, South Africa, and other South American markets.

One of our problem statements was to send WhatsApp campaigns to the Nigerian audience to understand how they would respond to our campaigns, so as to have incremental revenue based on the voucher or discount codes that we give to them, and also to measure the performance of the campaigns once we send them. The bottleneck situation was that we didn't have any campaign measurement capability. That's why we reached out to Segment team from Ireland. The idea was to get the data infrastructure ready, which is required for Segment to get the data streams. Once the data was set up, we were able to set up journeys based on our needs and pre-select the data identifiers, and we sent the campaigns using WhatsApp setup from Segment.

For the campaigns, we touched base with the architect from Segment team. They were very helpful in setting up a data infrastructure via regular sessions with our data architecture team and product team. We identified approximately 25 data fields which would be useful to select the primary and secondary characteristics for the campaigns. This engagement lasted almost two to three months. Once the data was set up and signed off by the high authorities, we tested the setup. Once we were confident in the setup, we then collaboratively worked with the marketing team to set up the campaigns and set up WhatsApp campaigns to start with. The performance showed that there was a 2% conversion rate on the campaigns that we sent, and we were happy with the prototype setup using campaign analytics of Segment.

In the initial rollout of the campaigns, we definitely saw a 2% conversion rate. The end-to-end return on average ad spend showed an improvement of 14%. That was good, especially because we were using Segment for the first time. This was near the industry standards.

The best features that Segment offers is the ability to connect with data of any type because we had a very complex data set which could be integrated into Segment systems. That was helpful. Another valuable feature was that Segment is part of Twilio and Twilio has its own suite of communication features such as WhatsApp, CRM, and others. We can integrate the communications of WhatsApp and email communications from Twilio to Segment. That was helpful again. Another feature of Segment was the scalability of Segment. We could replicate the same setup for other countries because initially, our main target was to send campaigns for international countries such as Argentina, Nigeria, and South Africa. The next country that we wanted to target was the United States. It was easy to replicate the setup for other countries.

Segment enhanced our capability to use third-party tools for enhancing our analytical capabilities because segmentation as well as the campaign measurement using Segment helped us also identify the AI component of measurement as well as sending campaigns. In a way it actually changed our mindset on how we absorb the data set keeping the long-term view objective in mind. It also helped us be more data-driven when it comes to campaign measurement.

One thing that was surprising and inconvenient at the same time was the risk of becoming too rigid with the data. The pre-selecting data identifiers was useful for the initial use cases, but as we got more mature with our use cases, it could become more restrictive because the business might later decide to introduce more data attributes or feeds. That means more dynamic segmentation requirement. Segment didn't have that good capability to be more dynamic with the inclusion or exclusion of data feeds.

Segment can be improved by centralizing the features that it has. I understand that it has gone through acquisitions, such as Twilio and Segment still operates as a separate system. If they can integrate the Twilio aspect of communication with Segment system itself, it will save our time when it comes to the infrastructure setup because we had to reach out to DevOps to have the right setup connected to our Twilio communications. It actually would save time for the clients of Twilio Segment to make sure that the systems are well integrated. It would also be good to have a good understanding of the inherent constraints because we remember that WhatsApp communications were not applicable for certain countries. That could be a part of the knowledge transfer in the onboarding discussions itself, so as to avoid any surprises in the downstream conversations. More support from WhatsApp conversations could be helpful. It could be possible that some countries don't allow multiple numbers. If we can have multiple numbers for the communications, that would be helpful to reach out to our target client base.

For the technical challenge, we can have more flexibility on the data feeds that Segment can incorporate from the client database because the data needs can change for any corporate, whether it be small, medium business, or large enterprise. If we can retrofit the data irrespective of some data fields being empty, that would be really helpful. It would also be good to have two or more data feeds input because it's difficult to get the data infrastructure done on the client side. It's better to have a customization of the data irrespective of the sources it came from. That would be really helpful for the data manipulation activities inherent to the segmentation.

The advice that I will give to anyone who is looking at Segment as a potential partner is to make sure that you have your problem statement ready. Ask the difficult questions to Segment about how the data will be used from a storage perspective and from a processing perspective. Are there any challenges inherent in their data which might not be understood by Segment system? Have the questions ready about the data infrastructure because the data is key to getting the right processing and the accurate processing of the data from Segment.

I have used Segment for almost 18 months in my previous company.

Segment is stable.

Segment scalability is good when it comes to geographic enablement but it is not scalable in terms of data. As I mentioned before, when different streams of data change, it might not incorporate once the data is set up. It should be flexible and robust enough when it comes to handling different streams of data once the product is set up.

Customer support is good. I won't say excellent, but they do respond and they have a dedicated customer support team.

This was the first time we used a third-party tool for AI segmentation as well as campaign measurement.

We worked with Segment team in 2023 and 2024.

Segment pricing depends upon the number of interactions and the number of campaigns sent. Irrespective of the great service and onboarding support, it is a little pricey. For example, we were still not ready with the data setup and we realized that the onboarding support hours were getting over and that was keeping our project at a standstill. We had to request the onboarding team for extra hours. After multiple requests, the request was granted for the free hours, but I think that should not be the case since the beginning of the project because the complexity of the project and all should be discussed beforehand and that should come from Segment because Segment is aware of the complexity inherent by working with different clients. Pricing could be a bit cheaper and more flexible and transparent in nature to avoid any confusion because the switching cost of any technology is very expensive for any organization and that can add to the frustration. Segment can do more competitive analysis when it comes to pricing its products.

I don't have specific numbers to share, but we did see improvement in the campaign's performance and ability of our organization to send campaigns without building the technology up front. We definitely saved money by not building the technology on our own. However, Segment is not that cheap. There's an investment of almost north of $1 million for more than three years contract. If Segment integration is very successful for any organization and the contract is long enough, then the ROI would be good. If not, then ROI might be a debatable issue.

We did evaluate other options and most of them were concentrated in US markets. We wanted to work for WhatsApp campaigns which could be worked in Argentina and South Africa. We zeroed down to Segment. I don't remember the names of other competitors we evaluated.

I have discussed all the challenges inherent in Segment that I have witnessed. The review rating for this assessment is 7.

My main use case for Segment involves taking information from a ticketing database and trying to reconcile it with customers that were coming from the website that were known to see if we could retarget and re-communicate with them over email.

I set up a reverse replication from their Redshift database and pulled all the customers and their primary key IDs into Segment with known email addresses. Any signals we got from the website where customers authenticated, we would send that email address as well so we could identify that customers from the ticketing database matched with website traffic. We put them into a segment inside of Segment, which we would push to the email platform Marketo to send customers emails based on their interaction on the website and the ticketing database.

We also enabled the feature of Segment called Engage where the marketing team could go in and create any other type of customer audience segments they wanted without needing IT help. Those segments would be pushed into their email marketing platform, Marketo, for them to further send any other types of emails they want, providing the marketing team with their own self-service.

The best features Segment offers are ease of use and lightweightness. You do not need overly complex technology capabilities to be able to program within Segment, unlike some other CDPs that are quite complicated. Segment is very simple.

A particular feature in Segment that stands out as especially user-friendly is that it was easy to set up the back-end data feeds and pull out these audiences without the need to create special data models, making it quite turnkey. I could pretty much set up a Segment platform within a week or two versus some other CDPs that take a lot longer.

Segment has positively impacted my organization by enabling my client to improve their marketing and understanding of their customers through the use of Segment without needing IT as a bottleneck.

Segment could enhance its identity resolution and identity management by providing more options to resolve IDs and possibly improving how it handles anonymous ID resolution since most cases are anonymous.

I chose a seven for Segment because while its lightweightness made it easy to implement, it also lacks a lot of other features. However, it is probably acceptable that it did not have more features, as Segment is fine for what it is capable of doing. They need to improve some features, particularly adding more AI enablement and AI discovery that were not in the product when I was working with it.

Regarding Segment's AI capabilities, I think its accuracy and reliability of output were lacking. It did not have any AI capabilities when I was working with it, and I do not know if Segment has added any more AI capabilities since then.

I have been using Segment on one client for the last nine months.

Segment is stable in my experience.

I did not see any issues with Segment's scalability.

The customer support for Segment was very important to us, as we had dedicated support that helped us troubleshoot at the beginning. Their professional services team worked hand-in-hand with us in setting everything up and digging into the use cases, which was important.

I would rate the customer support a nine.

I did not previously use a different solution.

I can pretty much set up a Segment platform within a week or two versus some other CDPs that take a lot longer.

Their professional services team worked hand-in-hand with us in setting everything up and digging into the use cases, which was important.

I have not seen a return on investment and cannot share metrics such as time saved, money saved, or fewer employees needed.

Based on my experience, I think Segment's pricing, setup cost, and licensing are fair. It is relatively cheap and very useful at that price point.

Before choosing Segment, I evaluated other options such as Hightouch and Adobe CDP.

My advice to others looking into using Segment is that if people want a lightweight CDP platform to collect information, validate through email in terms of identity resolution, and provide their marketing team with a lightweight tool to create audience segments, they should definitely look at Segment as a great tool to use. I would rate Segment a seven overall.