Twilio Segment

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.

I used Segment as a partner and a go-to-market partner for a customer data activation practice, and we were partnered together for two years very closely. Subsequent to that, I was still partnered with them, but we did less work together, and today Segment is a CDP that we would use, but my relationship with them is not as strong.

My main use case for Segment is for customer data activation.

I capture customer data, understand the patterns and trends going on, and test against those understandings in an omnichannel manner. The primary use case is for omnichannel orchestration and how we could personalize and make a better experience for an individual based on our understanding of who they were. Thus, the primary use case would have been personalization.

Segment is similar to most other CDPs in terms of its offerings, and where they differ.

At the time, one of the differentiators for Segment was the fact that they were part of Twilio and there were multiple technologies that would integrate, such as Auth and SendGrid, which created a whole omnichannel experience.

I really enjoyed working with Segment's partnerships team, as I did not implement Segment for myself but was a consultant that implemented it for others. Their partnerships group, when I was working closely with them, was very strong and very good. Plus there were strong market development funds to tap into, and we could jointly show up at conferences together and be thought leaders together, so I thought the program was good.

When we were at various conferences, we were able to show the specific use cases for customer data activation in retail and retail use cases and demonstrate how Segment connected to all the various Twilio components. This helped people see how it could come to life, how a call center could act on customer data, and how an individual could be authenticated so easily. Thus, us being able to demonstrate that and show the business value jointly worked really well.

Integration was hard. The integration of any CDP is hard, and the go-to-market messaging is that it is easy for any CDP, and it isn't. It would maybe be easy if a customer's data was in perfect order, but that is generally not the case.

The professional services were a little too simplistic. They had great people, but implementing a CDP is quite challenging, and I don't think that Segment teams thought of the whole architecture and the ecosystem and how Segment was one part of it. They became a little monomaniacal on what is the one use case and didn't recognize that data could live anywhere and needed to be harvested and transformed, leading to a hard time seeing the big picture. When we worked with customers who were using the professional services, they weren't always happy with the implementation because it was a little too simplistic and didn't live up to what the sales team said it could deliver.

I have been working in the marketing field for over 15 years.

I would rate customer service as 4 out of 10.

We deploy Segment for customers, so it is not specific to us. Therefore, it would be multiple different implementations.

I would say it is important to have a phased activation plan, think about your use cases, think about your data governance, and consider the environment that you are bringing Segment into. Having the technology is not a magic bullet, as there are many other factors that will make it successful. Really think about the customer experience you are trying to drive towards and work back from that. I would rate this review overall as a 7 out of 10.

My main use case for Segment is as a CDP to create a centralized view of the customer and help integrate the activation of audiences to downstream services.

A specific example of how I use Segment for this purpose is that we collect data from different data sources, and we then leverage the activation layer within Segment to populate data within Braze as the customer engagement platform. In practical terms, this could be somebody browsing our website or interacting with our loyalty system, and based on those user behaviors, we share those events and the data with those events with Braze, which is then used for personalized communications such as email, SMS, and push notifications.

That is the primary use case, and we have also implemented custom functions and transformation layers within Segment.

The best features Segment offers include the ease with which we can set up new data sources, the protocol structure they have for validating that the data is coming in the correct format, and the connection with back-end databases, mainly with Databricks where we both receive data via Databricks as well as push data to Databricks. The audience activation layer is also significant for our use case, and the ability to connect directly with the back-end database is very useful.

Segment's Databricks integration has helped our team significantly. In many cases from an audience standpoint, when building audiences, it often relates to customers as well as other data structures. With the database connection, I can create audiences, nested audiences, and other structures that connect customer data with many other types of data structures relating to loyalty, product, and purchases, and that has been a very powerful feature for us.

Segment has positively impacted my organization by making working with customer data a lot easier for us, as well as improving the speed at which we can ingest from new data sources and push to different activation layers. I have seen improvements with our time to market with Segment, which has improved dramatically due to the ease of use, allowing us to handle many use cases that were sidelined for years and now can be completed in days to weeks instead of months to years.

I think Segment can improve in terms of identity management, possibly in some of the protocols by adding more flexibility. I also see the need for more free-form searches such as regex or conditions that are not case-sensitive, which would be tremendously helpful.

I have been using Segment for about two years.

Segment has been stable in my experience.

The scalability of Segment for my organization's needs is very good.

Segment's scalability can be measured in terms of transaction volume, which it handled very well for us in terms of millions of records, and in terms of new use cases and functionality, we managed to find workarounds across various use cases by leveraging their functions and different activation layers, showing that from a scalable perspective, it performs quite well.

I have had experiences with customer support from Segment, particularly when we were learning the product and expanding our use cases. Their support team was extremely helpful in resolving issues, mainly when we struggled with the Databricks connectivity.

I did come from a previous system before Segment, but I am currently unable to recall what the system was, and it just was not adhering to all the needs we had.

The integration process with Segment was generally straightforward, although we faced a couple of challenges with Databricks, which I believe were due to corporate policies implemented within our Databricks environment that created the complexities.

I have seen a return on investment with Segment, primarily around time saved, and it also means we have had many more business-related users able to leverage it rather than needing development.

Before choosing Segment, we evaluated other options, including Azure Data, which was the next major competitor we considered, as well as Adobe Experience Platform.

For the basic functionality, the learning curve for new users getting started with Segment is very easy.

I find Segment relatively flexible within the defined functionality, and the ability to add custom functions helps extend that considerably.

Segment's documentation and resources provided are helpful.

Segment integrates very well with all our other tools.

I appreciate the simplicity in terms of getting into and leveraging Segment. Compared to some other CDPs, those options felt much more complex to reach the same stage, which is why I valued the simplicity of Segment.

I advise others looking into using Segment to understand who their intended users of the data are, what processing they want to conduct on that data, and how clean their input versus output will be. I would rate this product an eight out of ten.