- In August the team in charge of AWS’ marketing software Pinpoint pitched the idea of buying HubSpot.
- The AWS business-applications team has struggled to grow for years despite launching many products.
- The main AWS infrastructure business is getting commoditized, and competitors are catching up.
In August, the team in charge of Amazon‘s marketing software Pinpoint wrote a stinging review of its own product in an internal planning document that was presented to leaders of the Amazon Web Services cloud business.
The team concluded its brand was weak and the product was “too difficult” to use, according to the document, which Insider reviewed. None of the top analysts included Pinpoint in their reviews, and most customers were unaware of the software’s specific capabilities. One customer, Hulu, used Pinpoint only for simple tasks and relied on other software for more complex marketing campaigns, the document said.
To catch up, the Pinpoint team recommended buying a rival. The list included “marketer-friendly” startups, such as Braze, CleverTap, or Resulticks. They also suggested bigger names, such as Hootsuite or Sprinklr. And if AWS wanted to be really ambitious, it should acquire HubSpot, a marketing-software company with a public market value of $38 billion.
“Most customers do not consider Pinpoint as a marketing tool and are confused by Pinpoint’s product positioning,” the Pinpoint team wrote in its presentation. “To truly go big in Marketing Cloud, we recommend considering HubSpot.” Shares of that company rose as much as 2.6% on Friday, while most of the stock market slumped.
The damning assessment is emblematic of the challenges Amazon faces in a key part of the cloud market: business applications, broadly called software as a service, or SaaS. While AWS leads in infrastructure — computing power and storage services that power other applications — it lags far behind in the lucrative software segments higher up the technology stack.
Insider spoke with current and former AWS employees about the company’s SaaS struggles. They asked not to be identified because they were not authorized to speak publicly. They pointed to a deeply rooted culture, born out of success in the infrastructure business, that often serves as an impediment to building new SaaS offerings. Amazon CEO Andy Jassy‘s reluctance to make big acquisitions, and long-held desire to build products in-house, have also contributed to the setbacks, these people said.
“Jassy wasn’t interested if he didn’t feel like it was a sure-fire thing — that’s impossible in SaaS,” one of the people said. “Amazon executives have no idea what a modern SaaS application looks like.”
An AWS representative said “the premise of this article is misleading and inaccurate.”
AWS is still growing strongly. In its most recent quarter, it generated $16 billion in revenue, up 39% from a year earlier, the highest growth rate since early 2019. Some of its applications, such as the call-center software Amazon Connect, have seen steady customer uptake, people familiar with the business said. The Amazon representative also highlighted Amazon WorkSpaces and Amazon AppStream, which support tens of thousands of active customers.
“We have several applications that are very large successes,” the representative wrote in a statement. “Others are earlier in their journey, but we continue to believe they have meaningful potential.”
The stakes are high for Amazon to snag more share of the SaaS field. The market for business applications, including everything from Microsoft Office 365 and Salesforce sales and marketing software to
At the same time, AWS’ bread-and-butter infrastructure services are becoming commoditized, with its competitors Microsoft and Google catching up — in part by bundling with their popular software offerings, like MS Word and Dynamics or Google Docs and Gmail.
“Amazon’s had only limited success in finished applications,” said Matt McIlwain, an investor at Madrona Ventures, a venture-capital firm that has backed many SaaS startups. “What’s challenging is their two biggest competitors on the infrastructure and middleware front, who are Microsoft and Google, have both long-standing, proven application successes.”
AWS is expected to make several new announcements in the applications space at its annual re:Invent conference next week, according to The Wall Street Journal. AWS CEO Adam Selipsky is set to make his first appearance since replacing Jassy, who was promoted to Amazon CEO earlier this year.
Amazon’s view when it comes to new projects is to start small and invest more only when there are clear signs of growth potential. It’s a strategy the company has adhered to since its founding.
But in the SaaS space, where huge upfront costs are typically required across sales, marketing, and product development, Amazon’s playbook often served as a hindrance, people familiar with the AWS applications team said. It’s not uncommon for business applications to require years of development and sales cycles before taking off. For instance, Zoom, one of the most viral SaaS apps ever made, had to spend more than half of revenue on sales and marketing in the run-up to its 2019 initial public offering.
AWS leadership has had a hard time accepting this line of thinking, these people said. Part of that was due to AWS’ success in the infrastructure space, where it didn’t have to follow the conventional software playbook. AWS faced little competition during the first seven years of its launch because more established enterprise software companies didn’t take Amazon seriously at first, sparing it from having to spend excessively to gain an edge. Amazon’s founder, Jeff Bezos, famously called the experience the “greatest piece of business luck in the history of business.”
That created a Catch-22 situation for AWS’ applications team. To grow the product, the team needed more resources. But to get the resources, they needed to show results. Instead, one person said, most of the money at AWS went into infrastructure projects because “they are the ones who pay the bills.”
“The attitude was always, ‘We want to see traction first,'” another person said. “It’s really hard to compete for resources when you don’t see traction quickly.”
As a result, the team would languish with subpar products that wouldn’t even be used internally, despite having launched an array of business apps across word processing, email, and messaging, to name a few. There was no serious marketing push either.
“We did not broadly market them,” one person from the marketing team said. “The products just weren’t good. Why would a customer want to pay for them?”
To circumvent this problem, AWS’ applications team came up with an out-of-box idea, internally called “Rebel Alliance,” as Insider previously reported. Instead of building everything from scratch, AWS wanted to partner with popular app makers, like Dropbox,
, and Smartsheet, to create a bundle of business applications. The goal was to simplify the purchasing process for companies that prefer using a variety of apps rather than a suite of solutions from a single vendor, like Microsoft.
Humans versus machines
Another challenge was AWS’ lack of understanding of what’s important in a SaaS product, these people said.
SaaS applications require a different skill set than the infrastructure business. That’s because business applications target human end users, who need to find the software useful at work. In contrast, cloud-infrastructure offerings often focus on the efficiency of computer servers and other equipment in far-flung data centers.
“Infrastructure talks to machines, whereas applications work for humans,” one person on the applications team said. “Machines are much more predictable. Building apps is much more nuanced.”
That lack of creativity shows up in some of AWS’ SaaS products, the people said. For instance, the team at QuickSight, AWS’ data-visualization software, spends most of its time establishing “feature-parity” with its biggest competitor, Tableau, instead of coming up with more distinct features, according to one person on the team.
For Honeycode, a service that helps nondevelopers easily build applications, the difference in product philosophy led to a tension between the former vice president Adam Bosworth and Jassy, according to people familiar with the matter.
The Honeycode team kept missing deadlines and pushing the launch date because of small details, like shortening latency by 40 milliseconds, one of the people said. While the product team saw this as an important factor in meeting its high performance bar, Jassy grew frustrated with the delays, and ballooning costs that significantly exceeding initial projections, two people said. Honeycode eventually launched last year, but sales so far have been disappointing, these people added.
One person said every meeting was about the 800-pound gorilla in the room: Microsoft. AWS’ applications team was almost obsessed with its cross-town rival because of the dominance of Office in the productivity space, this person said. The Rebel Alliance idea, for instance, came out of those meetings because AWS suspects it may be too late to build individual applications on its own at this point.
“Amazon says it’s not competitor-focused, but every meeting was about Microsoft, Microsoft, Microsoft,” the person said. “The problem is Jassy hasn’t thrown enough money at the business.”
Unwillingness to buy
One recent setback in AWS’ applications team was the sudden departure of its vice president Larry Augustin, a high-profile hire from 2019. Less than two years into the job, Augustin left in part because of differences with Jassy, according to people familiar with the matter. His resignation led to a massive exodus in the applications team, including some of Augustin’s top deputies.
The tension between the two had a lot to do with resource allocation. Augustin constantly asked for more investment and headcount, as well as more acquisitions. Jassy took a more cautious approach, dismissing most of the ambitious ideas Augustin came up with, these people said.
Some of the acquisition targets AWS considered in recent years include Slack, Zoom, and even the encrypted messaging app Signal, as Insider previously reported. Those ideas never materialized, largely because of AWS’ culture of building in-house rather than buying for growth. A big concern, one person said, was whether AWS could add any unique value to acquired SaaS businesses, and vice versa.
“There was definitely tension as to why Amazon would make this asset better,” a person involved in AWS’ software-acquisition debates told Insider. “We’d ask, ‘Why do we think we’ll be more successful selling Slack than Slack is?’ — and nobody made the business case.” Salesforce ended up buying Slack for $28 billion.
Frustration grew within the team because Amazon’s retail division was making much more aggressive bets. In 2017, the company acquired Whole Foods for $13.7 billion, and the next year it spent another $1 billion to buy the pharmacy startup PillPack. AWS, on the other hand, has been more tempered with its acquisition strategy, mostly picking up smaller startups. AWS’ biggest deal to date is the $350 million purchase of the chipmaker Annapurna Labs in 2015, which helped mostly its infrastructure business, not its SaaS products.
It’s a decidedly different approach from that of Amazon’s biggest software competitors, which often rely on large acquisitions. Microsoft grew successful services, like PowerPoint and Excel, through acquisitions, and it has been an active buyer for years. Google built the foundational blocks for G Suite after buying numerous startups. Salesforce and Oracle go on massive acquisition sprees every few years. Selipsky returned to AWS in part because Salesforce bought Tableau, where he was CEO, for almost $16 billion in 2019.
AWS may be aware of this potential hole in its SaaS strategy. In June, it bought the encrypted messaging service Wickr, a move that could accelerate its share of cloud apps in the public sector and give even stronger end-to-end encryption. It’s a small step toward boosting AWS’ application offerings, but some say it may be too little too late.
“AWS is really frugal when it comes to M&A,” one person said. “They missed out on really big opportunities.”
“While we will continue to pursue mergers and acquisitions when we have the right fit on product, team, and valuation (as we recently did with Wickr), we have confidence in our ability to invent and will pursue a long term approach as most applications are reinvented over time,” the AWS representative said.
Do you work at Amazon? Contact the reporter Eugene Kim via the encrypted messaging apps Signal or Telegram (+1-650-942-3061) or email ([email protected]). Reach out using a nonwork device. Check out Insider’s source guide for other tips on sharing information securely.