Customers are the center of any business decisions, but in order to make any well-informed decision while reducing risks and uncertainty, it all comes back to data. They're at the root of customer experience management, the heaven of performance managers, the promised land of growth professionals, the fairyland of a product manager - well, you got it.
For every interaction that someone has with your brand (an ad click, visit in your website, an action in your product), they start a journey of multiple touch-points that generates hundreds of signals. What if you had the infrastructure to collect and understand this data, build models, and develop applications from them to drive a better customer experience?
And that's where Segment comes in. The IT company founded in September of 2011 provides customer data infrastructure that gives businesses the ability to collect detailed data about their customers, then deliver customized experiences based on what they have learned about them. It's all about customer experience management taken to another level.
In less than ten years, the company founded by 4 college students has raised more than $280M in funding over 8 rounds collecting now 15 investors. The latest funding was raised in January 2020 from a Secondary Market round. Segment has a post-money valuation in the range of $1B - $10B as of April 2019, according to PrivCo.
Fastly adopted by startups when they were exposed to the market, now they count more than 20.000 clients from different sizes, segments, and countries, including enterprise companies like Intuit, Atlassian, and Gap, who are using their platform to better understand their users and make customer-centric decisions.
By the end of 2020, big news took the market by surprise. Segment got acquired by Twilio ($TWLO), the public cloud communication provider, for $3.2 billion in stock in November of 2020. But what might surprise some at first, turned out to be a perfect match. With Segment providing customer data and Twilio being a cloud communications platform, these two companies combined help businesses to not only understand their customers better but also act based on data.
But in a world of heavily funded and popular analytics tools like KISSmetrics, Google Analytics, and Mixpanel, how could they possibly innovate? If you want to find out more about the ups and downs in Segment's trajectory so far, stay with us.
In this Growth Study, we'll look at:
- Segment 1.0: Failures happen
- Segment 2.0: What Product-Market Fit feels like
- Segment 3.0: Go-to-Market growth
- Segment joins forces with Twilio
- What future holds for Segment
Segment is a customer data platform (CDP) with the purpose of empowering every team with access to data. With Segment, marketing teams can deliver personalized and real-time customer experiences, product teams are enabled to make data-driven decisions, and help engineers to collect and manage data using one single platform.
Founded in 2011, Segment became one of the most essential tools for teams that rely on data to make decisions. Their technology provides data infrastructure to businesses that put their customers first, making it easier for them to collect data on different touchpoints, unifying them so they're sent down the stream to all tools used by marketing teams, product teams, finance teams, etc.
Helping more than 20.000 businesses from different segments, industries, and distributed over more than 70 countries, the API-based tool is revolutionizing the way companies send every online data they collect to different analytics or marketing tools, going further and helping them to combine that data with customer information from services like Stripe, Salesforce, SendGrid, and Zendesk.
With Segment, all the analytics data generated by your company is no longer fanned out. Instead, this information can be now collected by analytics services you use and store in a data warehouse. Based on this, companies that are trying to achieve a common understanding of their users become able to make customer-centric decisions.
Peter Reinhardt, co-founder, and CEO of Segment said:
Today, you have a mobile app, a website, email, push notifications, the call center, and so on and so forth. These are different siloed communication channels into the company. The result is that experiences used to be very personal, like maybe a bank branch manager who actually knew you as a person. Now, the bank or whoever actually has no idea who you are, because the data that happens on each of those channels and stays in those tools.
Let's say your marketing team is using Sendgrid or Mailchimp for their email marketing. You are also using Hubspot for your landing pages and Stripe as your payment gateway. So you end up having 3 different sources of data that need to be collected and interpreted to conclude how it is performing. Segment enables you to plug your entire stack of tech tools (Sources), creating a single source of truth and making it easier to both analyze the results or send the data elsewhere (Destination).
For Che Horder, Data Science Leader at Facebook and former Director of Data at Instacart, with Sources his team had a more complete understanding of their funnel, what was working, and the impact of their support team on revenue. Based on this, they were gathering new insights about their customer experience.
As a tool that allows companies to deliver the needs of their customers, at Segment they also put their customers' needs first. Proof of that was when they added BigQuery, Google’s Data Warehousing Service, to their integrations catalog.
At that time, 2016 to be more specific, Segment already supported other warehousing products, like Postgres and Amazon's Redshift, but they also saw an upsurge in demand for BigQuery. Although not all of their clients used their warehousing options, 'almost 100%' of their enterprise customers did, and not to mention a significant usage on the self-service side. But instead of just launching the new integration, they ran a beta test.
They assumed that their data warehousing customers were split 50-50 between Postgres and Redshift, but after one month of running the test, BigQuery had already claimed 10% of the total. Besides the early adoption, other associated advantages related to BigQuery were the fact it can be easier to use and it could turn their offer more appealing to smaller businesses.
In fact, adding integrations to their stack was one of Segment’s first growth strategies. Now they count with hundreds of integrations divided into more than 30 categories that can power your toll with the data they need, including FullStory, Mode Analytics, and Indicative.
To summarize and get straight to business, Segment can be explained as a technology startup that allows companies to pick and choose which third-party services to integrate into their data pipelines, pulling customer data from data and marketing analytics tools they use into another, by way of APIs, and use their platform to centralize this movement of customer data, used for marketing, growth or other applications that benefit from customer identification and segmentation.
Segment 1.0: Failures happen
It all started in 2010 when three college kids from MIT, Peter Reinhardt, Calvin French-Owen, and Ilya Volodarsky living in the same dorm, started reading essays by Y Combinator and started taking a class together at MIT called "The Founder's Journey". At that time they didn't have plans to become entrepreneurs but it was something that they would consider exploring.
Right In their first class, Adam Smith, MIT alumni, and founder of two successful companies, Xobni (email software startup), and Kite (machine learning startup), gave a guest talk. Amazed by the story of someone who had attended the same classes or walked through the same dorms, there was only one option: invite Adam for a beer at their dorm. And Adam accepted. It was the first lesson learned by the three friends: don't be shy and if you want something, ask for it.
From that chat, they found out that founders are real people like you and me, so why not try? And that's how they began their entrepreneurship journey.
But success doesn't come easily. By the spring of 2011, they came up with the idea of creating a "classroom lecture tool", allowing college students to flag points of confusion to professors during their classes. Segment was born (at least that's what they thought) and they promptly applied to Y Combinator, an American seed money startup accelerator.
They were accepted and now they had $600K in funding to launch their idea. What a big moment. With a year left at MIT, they decided to take a leave absence and moved to San Francisco. Peter Reinhardt was now the CEO; French-Owen, the CTO; Volodarsky, the president. And to complete the team, they invited Ian Storm Talyor to become the lead designer.
Little did they know that angel investment, team focus, and an idea aren't enough to succeed and that they were about to begin the second part of their journey: finding product-market fit.
Segment 2.0: What Product-Market Fit feels like
Finding PMF is not an easy task. In fact, 80% of new startups fail in finding product-market fit, meaning that 4 out 5 will fail and the one that doesn't fail will struggle to survive.
Reinhardt describes failing to find PMF as a bottomless emotional freefall that turns into an obsession. From his experience at Segment, he became an enthusiast of the idea that while you stretch your startup runaway, it is important you test more than one idea to get more shots at your final goal.
When they deployed the initial idea into classrooms in Berkley, professors showed excitement with the idea, and, for them, it was a clear sign they had achieved PMF. But when they went to the back of one of the classrooms to observe how the students were using the application, they discovered something entirely different.
They found out that the students weren't paying attention to the class and, therefore, weren't using their product. Instead, they were connected to eBay, playing Solitaire or scrolling down their Facebook feed. Everything but paying attention to what the professor was saying, and interacting with their tool.
That's when they learned their second lesson: if someone likes the idea of your product, this doesn't necessarily mean they will use it, or use it in the way you expect - or in other words, they did not find PMF just yet. Students weren't using their product because as soon as they were connected to the Internet, they stopped paying attention to the class and started surfing on the web.
These findings came as a bucket of cold water. Soon they called back to investors explaining the situation, and to their surprise, investors had predicted that their idea wouldn't work out, but they had higher expectations for the team itself. Even if the product-idea was not there yet, the team was right and ready - so thought the investors, who all stayed but two.
It was the fuel they needed to roll up their sleeves. They spent one year trying to build a web analytics tool so they could understand what was happening with students during class. If they had done this before, they would know that they were disengaged.
As soon as they started searching for the best tool to use, they knew they had three options: Google Analytics, KISSmetrics, and Mixpanel. But which one was better for them?
As proud developers, to answer this question, they decided to build their own piece of code that could take data from their website and distribute it to those three tools.
They began to work nonstop between fall 2011 and winter 2012 to create a new product for Segment, the one that would eventually become what we know today as Segment: a web analytics tool for companies to better understand their customers' behavior.
Following the lean startup playbook, they "got out of the building" and went out to ask occasional customers if they were interested in an analytics tool. From talking to them, they soon realized that people weren't interested in one more analytics tool. In fact, their biggest issue was having the right report or wish they had a specific dashboard.
The importance of getting feedback and iterating the product didn't stop there. From collecting customers' feedback they came up with new ideas and launched two more products:
Warehouses - to help customers collect and store 1st. party customer data;
Sources - to help companies analyze data generated by Warehouses
From this experience, one thing remained clear for the group: you only get this sort of clarity by talking to your customers and not by trying to be intuitive.
Running out of time and resources, they had to work on something they could test quickly. Almost like that experiment, you run by the end of the quarter that is Easy to implement and you're Confident that will have a great Impact on your strategy. But, it was not an easy test. So, what did they do, then?
For Taylor, the product was ready, they just needed to adjust their programming tool, transforming it into a simple API that would be used to collect analytics data from any platform. Reinhardt disagreed. At first, the idea wasn’t well accepted by Reinhardt, but as soon as he put more thought into the idea, he decided that they should run a test to validate Taylor's idea.
With their product ready, they launched Segment on Hackers News, a developer community run by Y Combinator, with a landing page that explained the value of their open-source library, and an email signup form to see if they were faded to success or failure. And Reinhardt was proven wrong, soon enough the launch was a huge success to the point where people were demanding (yes, demanding) access to it.
During his AMA (Ask Me Anything) with GrowthHackers, Reinhardt remembered:
The launch was a pretty defining moment for me personally since it was such stark proof that my intuition could be completely wrong in the real world,
And that's what PMF feels like. Rather than just surviving, investors will be interested in your product, people want to work for you, customers want to help you to improve, every little thing in your business goes crazy. That's what PMF feels like and from now on you're no longer focusing on PMF, now is time for Go-to-Market (GTM) growth, transitioning from early-stage to growth stage.
In 2020, during the Slush Technology Conference in Helsinki, Reinhardt was one of the speakers and with his talk "Segment's Bumpy Road to Product-Market Fit", he advised attendees how to avoid falling for the same pitfalls when trying to be successful and reach product-market fit.
And he added:
There is no learning encoded in failure but there is learning encoded in the success of knowing what it does feel like to find PMF.
Six months after launching the open-source library, Reinhardt said in an interview that at the beginning they had a good bit of traction, meaning some customers using the product for free and no revenue for them. And it was because of their customers' reviews and that initial traction that they were able to raise a $2M seed-extension round from Kleiner Perkins and E.ventures.
It was no surprise that Segment was initially adopted by startups. It was back in 2013 that they acquired their first customer, TheKnot.com, a wedding website. The negotiations happened between Reinhardt and their champion at that time, John.
At first, they started experimenting with Segment on some smaller apps they had and ended up falling in love with the product. After numerous phone conversations, leading to a very slow sales process, John continued to test Segment pushing it into more and more products. When their data volume got substantial, falling into Segment's business purpose to help companies to better manage their customer data, he decided that it was time to negotiate a contract. Reinhardt remembers that because he was terrible as a sales rep, John got a pretty awesome deal. It wasn't the perfect negotiation, but still, it was a win!
Another early adopter of Segment was TravelPerk, a travel management startup, that uses the platform to unify its data collection instead of letting different departments use different metrics tools and different data warehouses but still not being able to answer simple questions like the conversion rate from search to purchases.
With one year of business rolling, they crossed one million dollars in revenue while still growing super fast. Peter remembered that in the six months leading up to the Series A, they went from 0 to 1 million in revenue, and they were on track to end that year adding more than 1.5 million in MRR. If they had questions about standing out in a very crowded market, now they didn't.
As they were growing and acquiring more and more startup customers, they saw a big opportunity in selling the technology to large enterprises. That's when IBM comes in, the traditional software and hardware company, using Segment to systematize data collection across its cloud services so it could be processed and analyzed by its marketing tools from different segments, industries, and countries (more than 70 countries).
In 2017, Segment raised $64 million in Series C funding, with Y Combinator Continuity Fund and GV leading the round. The continuity was the $700 million fund led by the former Twitter COO, Ali Rowgahni. It selectively writes for YC startups as they raise bigger rounds. When Rowgahni joined the board of directors, he described Segment as one of the most successful YC Combinator companies. This round brought Segment's total funding to $109 million, adding to previous rounds joined by Accel, Thrive Capital, SV Angel, and NEA.
By 2019, Segment raised $284M in venture capital funding and tons of millions of dollars in ARR, said Reinhardt to San Francisco Business Times. Until February of 2020, they had raised a little less than $300M in VC serving more than 20K customers in more than 70 countries all over the globe.
Until February of 2020, they had raised a little less than $300M in VC serving more than 20K customers all over the globe.
Segment 3.0 - Go-to-Market Growth
Once you reach PMF, you're not shielded from failure. Is time to aim for GTM growth.
From the early days, having a growth mindset was part of Segment's DNA. Launching their product as an open-source library for the dev community, or iterating the product as they spoke to customers were indications that the company was all about experimentation.
During their trajectory, we've selected some of their experiments to share:
1. Change the plan, but not the price
One of the challenges faced by Segment in 2015 was when their growth rate slowed a little bit. They were coming from an impressive growth trajectory when over the first years they went from zero to $2.5M, and then $2.5M to 10M, and then $10M to 20M. But this last year didn't happen smoothly. Used to grow at higher rates, after reaching $10M they only went to $12.5 million in the first half of the year. It was a clear sign that something was off.
What happened was that they had a pricing and packaging issue. Back then, Segment had packaged separately two different parts of their product, and people were churning off one of the packages because they found out that there was a cheaper way to make things work, they just needed to bring a 3rd. party tool.
Going on this way, their churn rate was spiking in the first year and about to become a disaster for their revenue. Their solution was to change their prices and packages by transforming the two separate packages into a single one and then changing their self-service pricing relative to their sales list price for larger contracts.
The experiments proved to be right. Churn started to decrease and customer acquisition was spiking.
2. CX insights fueling product improvements
Customer success requests and product development and partnership decisions are taken seriously at Segment.
The engineering team works closely with the support and customer service team - as well as the business development team - where they flag common questions, issues, suggestions, requests from its users and make sure the information gets to the right person.
Depending on the topic and the volume of comments/requests, they can become help articles or docs improvements that the success team write in big chunks of time they call “Docs Crushing”.Integration teams also always working to enhance their existing integrations and adding new commonly requested integrations.
Another team that works on customers' requests is the product development team. Usually, a new project comes when a question/issue becomes so prevalent that it needs to be solved in a bigger way than just a link to a help article or new integration. It is important to mention that once the new project is kicked off, they make sure to include a customer-success team member in the product development, planning, and building process so they stay close to their customers' needs. Plus lots of alpha and beta versions with a small subset of customers happen to get things rolling.
3. Personalization is key
Their marketing team is also always experimenting. Back in 2016 when doing paid media acquisition tests, they came to the conclusion that, for them, retargeting works while search doesn't.
Back then, from their early experiments with retargeting paid campaigns, they realized that to achieve their goals, they would need to optimize their landing pages. With LPs optimized, they customized ads based on what page that visitor has visited.
For example, if someone visited Docs’ page, they would start getting more API-related messaging. If someone visited the pricing page, they would get more free trial messaging. Learnings from this experiment showed an increase in their conversions while they were retargeting across the web, Facebook, and Twitter. With Twitter showing to be more profitable with the highest ROI.
On the other hand, the search didn’t work for them. Diana Smith, Director of Brand and Product Marketing at Twilio, said that this happened back then Segment was a new concept and people weren’t searching for it. But still, they had some success with the “KISSmetrics vs. Mixpanel” type of keywords, and branded search. But compared with the results they were having with retargeting campaigns, their efforts for making this strategy happen weren’t paying off. With traffic being so low, this couldn't become their focus.
4. Choosing the right target
Building and marketing a product for both technical and non-technical users with different motivations can be daunting, but Segment seems to have solved this 'problem' of serving multiple audiences.
They adopted the mentality that your marketing website has to choose a primary audience, and for them, their audience was Engineers, while all the sales process and supporting aspects of the website can address the desires of marketing, product, data, and C-level.
This strategy becomes crystal clear if you visit Segment’s main page where you'll find a code image right there and, as you dive deeper on their website, you’ll find explanations for other constituencies. Simple as that.
5. Maximize your voice
At GrowthHackers' AMA, Reinhardt remembered one time they were discussing user acquisition strategies and one of their investors said point-blank: "Guys. You know what your best marketing channels are going to be. They're the channels that are authentically Segment: developer evangelism and co-marketing with partners.".
A bit of external perspective and some time to let the idea sink in can sometimes help clarify what your core strengths are, and how you can leverage those to grow faster. And of course, this strategy was already apparent in their numbers: after word of mouth, their major acquisition channels are our open source work and co-marketing with partners!
The problem is that WOM is something that you can’t control. But their open-source work was and is something they can control. Building open source libraries like myth.io, metalsmith.io, nightmarejs.org, and github.com/segment.io/deku is an authentic way to help their core engineering audience and from their Google Analytics, they could actually see these referees appearing on the reports.
Jake Peterson, former VP Customer Success at Segment once said:
Our open-source projects have been huge for us in the developer community in 2 ways. Being genuinely open and helpful has been an awesome user acquisition and awareness channel. Also, we have amazing engineers on the team that share our love for the open-source community that may not have joined if there wasn't plenty of opportunities to explore, learn, and share outside of Segment's core business model.
And although they didn’t build these libraries for marketing purposes, they ended up discovering how they could share with the world amazing things that they’ve created internally. And it was also an opportunity to expose developers to Segment and some of them to decide to actually use them at work.
6. Back on track
When their team redid their tracking plan, they defined “Sent project data” as their ultimate funnel event to signify activation. A decision that was made using best practices from their success team with the help of their growth team. Measuring this event took a lot of their time and effort, but once it was working, soon enough they tested two ideas:
Redoing their billing emails. The process was broken but once they worked on it, they saw a large uptick in revenue by tightening up the entire billing process, a step in their funnel that was the most dysfunctional.
Redoing their activation emails: after long working journeys, now they know exactly how many visitors activate after they sign up. From now on they had a baseline to start running experiments, like redesigning activation emails and setup flow, to improve these conversions
7. Use the moment to your advantage
When Segment announced their Series A funding, they observed massive growth in terms of visitors. This result was expected since they took this opportunity to combine multiple marketing strategies, there were press releases, community outreach (Hacker News, etc.), and customer outreach around the launch.
To better use the momentum, they've also changed the domain from segment.io to segment.com and redesigned the entire site, making it easier for visitors to understand who they were and learn more about what they could do for them.
All these efforts culminated in a big pump in organic traffic and a spike in their conversion rates. This momentum, new visitors and customers, was then sustained by their launch of Segment SQL.
8. What's better for your business
Soon they had to make a strategic decision: charge the other side of their marketplace instead of/or in addition to charging end customers?
According to Reinhardt, early on they looked into an affiliate revenue model, but it was vanishingly small compared to the value they provide (and receive) with the end customer. Charging their partners would also dilute one of the biggest benefits they provide to customers, so they chose to be unbiased and therefore able to provide a balanced perspective.
Although Segment is a major source of leads for some of their partners, letting partners pay them would mean changing their core constituency in a way that would subvert some of their core value in the ecosystem.
Segment joins forces with Twilio
It was early November 2020 when the press announced the close of Twilio's acquisition of Segment where both companies are existing API-based tools but while one serves the purpose of building and running communication services, the other one provides customer data for better management.
Twilio, which trades at a market capitalization of more than $45 billion, went public in June 2016 on the strength of communications infrastructure allowing companies to track their visitors and customers through multiple channels, like text messages, voice messages, and video calls.
And now, clients of Segment can handle the growing number of customer information that arrives every hour and had previously been stuck in different compartments, making it difficult to have a single view of a customer and understand their path. With a more complete set of data on each customer, clients can now experience personalized services based on their behaviors and needs.
This acquisition by Twilio is a big step up for Segment in terms of scope and potential companies to target with its tools, especially with cloud companies booming during the pandemics. For Twilio, this acquisition is also a great opportunity to bring customers, opening the door to upselling. Is for sure a win-win situation.
Jeff Lawson, co-founder, and CEO of Twilio said:
Data silos destroy great customer experiences... Segment lets developers and companies break down those silos and build a complete picture of their customers. Combined with Twilio’s Customer Engagement Platform, we can create more personalized, timely, and impactful engagement across customer service, marketing, analytics, product, and sales. We are thrilled to welcome Segment to the Twilio team.
For Lawson, the biggest roadblock to better and personalized engagement between companies and customers has been that data has been locked in silos, and that's precisely the kind of problem that Segment has been attacking. Combining both companies is key to building great digital experiences.
We created Segment to help businesses set themselves apart in the digital age and deliver rich, connected customer experiences built on high-quality data. By joining forces and applying our customer data platform to Twilio’s engagement cloud, we’ll be able to make the entire customer experience seamless from end-to-end.
What the future holds for Segment?
Reinhardt hopes to see Segment turn into a broader platform where 3rd. parties can build products that tap into this data. By 2017 they had already reached an impressive mark of 200 built on top of Segments with 60 of them being created by partners and they didn't stop there.
He claims that clean and connected data should be top of mind for every company in 2021. If you have a team that is proactive with data, they'll lead to better customers' experiences culminating in better retention and revenue. If companies procrastinate, soon looming privacy regulation will become the fuel they need to take data practices more seriously, focusing more on 1st. party data.
I think the world that we see ourselves and all our customers moving towards is one where all of the data across all of their customer touchpoints is collected and synthesized into a modern customer data system of record, and that’s what we’re trying to build.
At Segment, the plan is to go one step further to link clean data to companies allowing them to harness the power of AI/ML applications and do things like automatically rank their content or get the right product in front of customers ready to purchase. Reinhardt bets that consumer goods companies that test with customers and iterate will take greater advantage of this improvement of AI/ML applications.
We find that the biggest thing blocking people from doing machine learning is actually having clean data that are formatted correctly for the input into the machine learning process...You talk to a data engineer, and they’ll say they spend 80% of their time ‘munging’ data and 20% of their time complaining about it. So what we’re trying to do is take away that 80% of the time munging the data. We’re working on a new beta part of the product that we’ll announce in the coming months that eliminates a lot of that munging.
Segment changed the way companies deal with their biggest assets, their clients. While we can't say exactly what the future holds for them, we know they will keep revolutionizing the way we deliver digital experiences.
With GrowthHackers Experiments, growth teams can integrate their experiments with Segment. Click here to learn more about our growth hacking management platform.
Don't forget to upvote this growth study if you liked this content and share with us in the comments your thoughts about Segment and if somehow they impacted your business.