Cutting Slack: When open source and team chat tools collide

Let the OSS Enterprise newsletter guide your open source journey! Sign up here.

In a 12-month period that has seen Salesforce commit $27.7 billion to buy Slack, Microsoft Teams skyrocket to 145 million users, and Asana hit the public markets, the past year has highlighted the growing importance of team chat and collaboration tools in an increasingly distributed workforce.

Another key trend accelerated by the pandemic has been the rise of open source. This is particularly true in the enterprise, judging by many accounts. Why? Well, as businesses have had to evolve and embrace digital transformation, open source code offers an easier conduit to scaling software, given that it saves companies the effort of starting from scratch often gives them more flexibility to create a differentiated product. Microsoft also went so far as to conclude that open source is now the accepted model for cross-company collaboration, as big technology companies often join forces to solve problems by investing resources in projects that benefit them all.

At the intersection of these two trends are open source team collaboration tools. There are any number of reasons an enterprise might decide to explore communication software that adheres to a more open philosophy. Companies that manage sensitive data, for example, might prefer to be in full control of all their information under a self-hosted option, not to mention the flexibility in terms of integrations and deployments.

Here we look at some of the open source “Slack alternatives” currently on the market and speak to the key movers and shakers behind them.

What matters most

Founded in 2011, Mattermost targets enterprises with various self-managed and hosted software-as-a-service (SaaS) options. The Palo Alto, California-based company has raised around $70 million since its inception and claims some major customers, including Samsung, SAP, Deloitte, Nasdaq, and BNP Paribas, with a typical “large enterprise deployment” of between 10,000 and 40,000 users.

Mattermost offers various plans covering most potential use cases, including its foundational free and open source Mattermost Team edition. It also offers a commercial self-managed free edition called Mattermost Enterprise Edition EO, which has the added ability to upgrade to more feature-rich paid versions. And if a customer on a hosted enterprise plan is concerned about their data becoming locked into commercial software, they can “downgrade” to the open source Mattermost Team edition without losing any data.

According to Mattermost CEO and cofounder Ian Tien, the most powerful benefit of operating what is known as an “open core” business model on top of an open source product is efficiency.

“Right now, the financial markets favor enterprise SaaS as a category, and you can see that through the validation multiples for the category,” Tien told VentureBeat, though he added that enterprise SaaS has two inherent challenges.

“First is that product differentiation only lasts six to nine months before competitors in the space have effectively rivaled your new features — it’s difficult to create a moat,” he said. “Second is that enterprise SaaS go-to-market strategies are commoditized — marketing and sales motions eventually all look the same. Everyone’s running ads on the same keywords, they use similar messaging, they go to the same conferences.”

The main problem this creates, according to Tien, is that as investors plow more money into enterprise SaaS startups, the cost of customer acquisition “goes through the roof.”

So how do open source software and open core business models address this, exactly? It all comes back to that key word — efficiency.

“Unlike enterprise SaaS, there’s not yet a lot of competition in open core markets,” Tien explained. “The vast majority of open source projects aren’t meant to become businesses, so when you build a venture-backed business in open source, you can often become the leading open source player in a category through engineering investment and sustain that position as you scale.”

This also has a knock-on effect on companies’ go-to-market strategy. It’s often developers and IT professionals who go searching for solutions to whatever problems they’re having, and open source is often top of mind for that target market. Businesses that have built a commercial offering on open source software are well positioned to lure these developers in and then tempt them at a later point with premium features. In other words, it’s often easier to get potential buyers’ attention with open source and help them realize its benefits without going through a long sales cycle.

“Open core companies can build offerings to address those needs, and it’s only after value is delivered that any conversation about commercial offerings appears,” Tien said.

Finally, it’s worth highlighting one of the defining attributes of open source software: community contributions. In the closed source realm, anyone is free to suggest or request features through forums or social media, but in the open source world it’s not only easier for end users to propose these features, they can build and implement these themselves.

“Some of our most popular, innovative features have been proposed or implemented by the community,” Tien said. “If Mattermost were an entirely proprietary product, we surely would have missed out some of the amazing ideas coming from our contributors.”

Rocket science

Founded out of Brazil in 2016, Rocket.Chat is a similar proposition, insofar as it offers various hosted and self-hosted tiers on top of its core open source chat platform.

More technically adept developer teams can also deploy the free open source “community” edition though Rocket.Chat. Chat sells additional features and services to companies that like the “open source” philosophy but lack the resources to do everything themselves.

“Some admins deploying community edition on-premises may lack particular technical configuration, operational expertise, or resources,” Rocket.Chat founder and CEO Gabriel Engel said. “In these cases, we provide certain supporting services on a subscription basis — bundled in our plans — that make their lives easier. But all of these services are optional, and the more technically capable users of the community edition will not need them from us.”

Rocket.Chat’s hosted SaaS offerings come into play for those who want to avoid the hassle of setup and maintaining their own servers.

When a company chooses a SaaS product, it typically has to accept that there will be some trade-offs in terms of the control it has over things like customizations and access to data.

“SaaS adoption requires a minimal level of acceptance that you’re no longer entirely in control of your corporate data and you must trust a third-party based on their — often constantly changing — terms,” Engel explained. “And in these situations, their capabilities and desires to ‘build new features’ and ‘new integrations’ typically diminish.”

SaaS products such as Rocket.Chat, however, combine the best of both worlds — the convenience of SaaS with the flexibility of open source and the ability to customize workspaces with different user permissions, third-party integrations, and centralized data the customer controls.

“Rocket.Chat has engineered a solution for even those cases where companies are willing to bet on SaaS and still want the complete flexibility to create new features and customizations,” Engel said.

As with Mattermost, Engel cites its community of users and developers as the main benefit of its business model. This creates a sort of “flywheel” effect, in which the community improves the product, which in turn benefits the community and its paying customers. This virtuous cycle encourages community participation to improve the product even further.

“Open source is incredible for businesses’ adoption — it removes purchase barriers, allowing customers to try the product and customize it for their business needs,” Engel said. “In case a feature or integration is missing, the customer can develop it themselves, contribute it back to the main, or publish it to our marketplace.”

Engel added that most of its largest customers are companies that started out by deploying its free and open source (FOSS) edition on-premises, only to ramp things up later.

“As their internal adoption grows and the company becomes more dependent on Rocket.Chat as the core communications and collaboration hub, they transition into our customers to get support for their projects and ensure the longevity and sustainability of our platform,” Engel said.

Rocket.Chat raised $19 million in funding just a few months back, nabbing some notable institutional U.S. investors, including Greycroft and NEA. And it claims big-name clients such as Continental, Lockheed Martin, and Seeking Alpha, an online community and publication covering financial markets.

Seeking Alpha launched the Seeking Alpha Marketplace back in 2015 to help investment experts sell their expertise to the wider world. As real-time chat is integral to such a service, the company initially developed and hosted its own chat client in-house, but this turned out to be a bad idea.

“We didn’t have the bandwidth as an organization to develop it properly at that time,” noted Jonathan Liss, senior product manager for the Seeking Alpha Marketplace. “And so the feature set really wasn’t competitive for anyone that had used a ‘real’ chat product. More importantly, every end user thinks about chat features, but from the development standpoint, the backend integration is the complicated part. A lack of scalability on the server side can lead to constant CPU spikes and crashes — and this is exactly what happened to us.”

After a year or so, Seeking Alpha decided to seek out an alternative product and eventually landed at Rocket.Chat’s door.

Using an open source chat tool for its marketplace enabled Seeking Alpha to do things it otherwise would not have been able to do. Scalability was one issue — Slack charges on a monthly per-user basis, something that could prove prohibitively expensive for a product with no tangible upper limit on the number of users.

Seeking Alpha also explored SendBird, a company that creates a chat API for businesses to integrate into their apps. However, this would have entailed developing its own chat front-end for each platform across desktop and mobile, when the company needed something that was good to go.

At the time Seeking Alpha switched to Rocket.Chat, it was probably a little less feature-rich than the likes of Slack, but the features weren’t the main draw. The key factor was flexibility and the ability to create exactly what was needed.

“The [Rocket.Chat] team there was committed to working alongside us to make sure we ended up with any features we did need,” Liss said. “We needed really tight integration between our production database and Rocket.Chat’s.”

The Seeking Alpha Marketplace has around 180 individual communities, each centered on a specific investment vertical, with private chat rooms people can access through subscriptions. Seeking Alpha needed granular controls to, for example, cancel someone’s access immediately after their free trial expired.

“And because market circumstances change so quickly, it was crucial [that] subscribers were alerted throughout the trading day about key goings-on in chat across a variety of mediums,” Liss continued. “Our older subscribers generally prefer email or desktop notifications, whereas the under-50 crowd much prefers mobile or smartwatch push notifications.”

Another appealing feature was the ability to mute individual users. This is something likely to be more useful in a large anonymous community setting than an enterprise — which is why Slack doesn’t have such a feature — but it highlights some of the flexibility open source tools can provide.

“Having a robust chat platform is ultimately about facilitating the free exchange of ideas and information,” Liss said. “And this means giving each individual member of a chat room the power to control their own informational flow and, ultimately, maximize their signal-to-noise ratio.”

It’s worth noting that companies sometimes use open source chat tools for very specific use cases. Seeking Alpha hasn’t gone all in on Rocket.Chat — for standard, day-to-day internal communications, it still uses Slack. “It [Slack] makes more sense as an IT use case — and it’s what our IT team is familiar with,” Liss said.

This is a theme that permeates the open source chat and collaboration space. Mattermost’s Ian Tien said he often encounters companies using the free and open source version of Mattermost for very specific situations.

“We commonly see just the open source version being used in high-security organizations, where small teams want to work together rapidly and effectively,” Tien explained. “Sometimes this happens with small teams working in air-gapped networks, sometimes called ‘enclaves.’ Those use cases for a couple of dozen team members working in Mattermost are ideal for the open source edition.”

Tim Abbott, cofounder of open source team chat platform Zulip, notes that he often encounters companies that use Zulip across the board, except for their sales team, which continues to use Slack. “One reason why Salesforce wanted to buy Slack is that Slack was a better UI for Salesforce than the Salesforce website, which is painfully slow,” he said.

Speaking of Zulip …

Keeping context

Zulip was built for asynchronous group chat, one of the pillars of the pandemic-driven remote work movement, with clients including cloud and content delivery network (CDN) giant Akamai and Wikipedia parent organization the Wikimedia Foundation.

Zulip’s main selling point is its unique threading functionality, replete with search and configurable notifications. It is effectively positioned as the anti-Slack. “Our goal is to make the most productive collaboration software possible, not to copy Slack,” Abbott explained.

With Zulip, users subscribe to “streams” (a little like Slack channels, for want of a better analogy), with each message in the stream also attached to a topic.

So users logging into Zulip after a day offline — or based in different time zones — can click on a stream and choose the topics that are relevant to them, rather than having to read through hundreds of missed messages.

While Slack is undoubtedly great for certain things, it can become archaic when dozens or more employees — all based in different locations — are trying to hold conversations. Zulip promises to help specific conversations retain their context, even when the messages contained within those threads are sent hours or days apart.

Zulip has taken a rather circuitous path to where it is today. The company was founded in 2012 as a workplace messaging platform, but it was snapped up by Dropbox in 2014 before it had even launched to the public. Dropbox had intended to use Zulip as the foundation of a new team chat feature, but priorities shifted and Zulip’s creators ended up working on other parts of the main Dropbox product. Dropbox eventually shuttered many of the new products it had been working on.

Throughout all of this, Zulip’s original roster of beta customers were still using it. “All of our customers tried Slack at one point or another and switched back to using Zulip even though — at the time– it seemed like the product might not have a future,” Abbott explained.

Dropbox released Zulip under an open source license in 2015, which was the first time that it had been made available beyond early beta users. Dropbox relinquished all ownership of the platform and eventually donated the trademark to Kandra Labs, a company Abbott set up in 2016 to oversee and sustain Zulip’s development.

Kandra Labs has since launched several commercial products to monetize Zulip, including a hosted incarnation and an on-premises version that includes enterprise support. Similar to others in the space, Abbot said most companies elect to self-host for reasons beyond the financial incentive.

“The main benefit of hosting with Zulip Cloud over the open source version is that you get our operational team managing the service — we take care of backups, avoid downtime, support end users, and set up integrations with third-party services like Google, GitHub, and Twitter,” he said. “Most companies that choose to self-host do so because they have a compelling security/compliance reason to do so, not because they want to save money.”

On the SaaS side, Abbott notes that one of the main differences between proprietary and open source products is that the latter allows them to add functionality themselves.

“If there’s a feature a given customer really needs, but it isn’t an overall priority for Slack, there may be nothing the customer can do to get the feature added to Slack — at least, if they’re not big enough to demand it as part of a sale,” Abbott said. “In contrast, with Zulip’s model, the customer has the option of assigning one of their engineers to implement it and submit the changes to be merged into Zulip. This makes it possible for a SaaS customer to ensure an important feature or bug fix that might not be possible to implement via an API/integration.”

Unlike Mattermost or Rocket.Chat, Zulip has elected to remain self-sufficient by not taking on venture capital investment, an arrangement Abbott is all too familiar with, having backed some 100 startups himself as an angel investor.

“We have chosen not to take venture capital because doing so creates strong pressure to compromise values,” Abbott explained. “We plan to fund Zulip’s growth the old-fashioned way — through users paying for our product.”

Element in the room

Element is an open source instant messaging client built on Matrix, a decentralized, open standards communication protocol developed inside Amdocs by Matthew Hodgson and Amandine Le Pape in 2014.

“There is literally no center, just as the internet and the web have no center,” Hodgson told VentureBeat. “Matrix is more than an open source project — it’s a whole ecosystem for interoperable real-time communication. In fact, we think of Matrix as being the missing communication layer of the open web.”

To harness Matrix’s potential, Hodgson and Le Pape in 2017 formed a new standalone company called New Vector. It went on to develop a Matrix hosting service and a cross-platform Slack alternative called Riot that was built on Matrix. In 2018, the Matrix.org Foundation came to fruition to spearhead Matrix’s future development as a neutral not-for-profit entity. And then last year New Vector and its Riot app rebranded as Element.

It is worth stressing that Matrix isn’t a product in the sense that Mattermost, Rocket.Chat, or Zulip are. In many ways, Matrix is more akin to something like email, which is perhaps one of the best examples of a successful interoperable communication system to emerge in the internet age. With email, users don’t have to be with the same service providers to message each other (though that wasn’t always the case). This is pretty much the antithesis of where we’re at with modern internet-based communication tools, including VoIP (e.g. Skype or Google Hangouts) and messaging apps (e.g. WhatsApp or Signal), which promote locked-in ecosystems that prevent users from chatting freely to each other across services.

Because Element is built on Matrix, users are essentially participating in an open decentralized global network of 30 million users spread across 65,000 separate deployments run by different organizations. It’s “More like a phone network than a typical proprietary silo, like Teams or Slack,” Hodgson said.

This means users aren’t locked into the Element app and can switch to any other Matrix client. They also get interoperability so they can communicate with anyone else using a Matrix-enabled app, while a method called “bridging” enables it to support non-Matrix apps, including Telegram, Slack, WhatsApp, and open source software like Mattermost.

“Someone on Element can transparently message back and forth with someone on the other services, without the other party even really being aware that they are talking to someone who is really being teleported in from elsewhere,” Hodgson said.

Element hasn’t actually implemented all these bridges yet, but it does support Slack, Microsoft Teams, Telegram, Discord, IRC, and developer-focused chat platform Gitter, which Element acquired last year.

Another core promise behind Element is data sovereignty — letting businesses choose where and how their data is hosted. “This is a huge deal for nation-states, regulated businesses, or anyone who doesn’t want to hand all their data to a single big tech company and pray it doesn’t get breached,” Hodgson said.

A number of commercial companies have built products on top of Matrix, including Ericsson’s Contextual Communication Cloud, which is a fully managed service that lets enterprises easily integrate advanced communication and collaboration services into their applications. And French giant Thales launched an instant message service for businesses called Citadel.Team, which bears some similarities to Element.

Among Element’s functions — besides being a commercial company with backing from WordPress.com’s parent company Automattic — it serves as a poster child for what is possible with Matrix.

“The No. 1 reason we develop Element as open source is, pragmatically, to act as a catalyst for the growth of the open Matrix network,” Hodgson said. “The way to get a new technology to spread across the internet is clearly to release it for free as open source with a large organic community of developers to drive it forwards — just as the web owes much of its success to the distribution and communities around the open source Apache web server, or open source Mozilla and WebKit browsers.”