The formula for high tech wealth creation sounds simple: hire the most talented people, build a product that delivers world-beating value to customers, and keep adding new products with great growth potential.
That results in sustained fast growth and a soaring stock price that provide the resources to keep that success cycle spinning.
One company that has been following this prescription is cloud networking provider, Arista Networks. Between 2011 and 2020, its revenue grew at an average rate of 36.5% while its stock soared at a 29% average rate between its 2015 IPO and the end of 2020.
Despite recent growth that has slowed from that long-term rate, Arista’s stock is up 16% in 2021 — the S&P 500 has risen 14.3% during the period — after it reported expectations-beating results in the first quarter.
After speaking with Arista CEO Jayshree Ullal on May 6, I see five reasons that make Arista looks to me like a good long-term investment because it
- Hires and motivates world-class networking talent
- Drives growth in its current products while investing in future ones
- Empowers those closest to the customer to decide and act
- Sets and achieves ambitious goals
- Keeps processes from curdling into growth-impeding bureaucracy
(I have no financial interest in the securities mentioned in this post).
Getting Arista Off To a Strong Start
Arista sells hardware and software networking technology. Hyperscale data center operators such as Facebook
After a stellar career at Cisco Systems, Ullal joined Arista when it was a startup. As she said, “I joined the company in 2008 when it had 30 employees. [We built from a] foundation with three aspects: we are innovation driven, we don’t do me-too. and we are software as a service. I have written 100 blogs on the topic of the cloud and started before it was widely adopted. I was inspired by Google
The financial crisis of 2008 turned out to be a crisis and an opportunity for Arista. “Five years later that foundation is serving us and our customers well. We needed to prove it would work with customer driven use cases. In 2008 there was the financial crisis and Lehman Brothers, our customer, went out of business,” she said.
The good news was that the product Arista developed for Lehman helped its peer companies survive. “We had built a system for the finance vertical and it helped our other finance customers to survive because we had lower latency — nanoseconds versus milliseconds and a software and stack that was programmable. The economy crashed and the customers loved it,” Ullal told me.
Arista’s Expectations-Beating Q1 Report
Arista expects strong growth in the second quarter of 2021, CFO Ita Brennan forecast second-quarter revenue of between $675 million and $695 million, according to its Q1 2021 Earnings Call Transcript. The midpoint of that range is 26.6% above 2020’s Q2 revenue and $35 million more than analysts had forecast.
Ullal credited AI with helping to boost its growth — with the quarter being the sixth in a row that Arista has exceeded analysts’ expectations. As she said in a statement, “Arista begins the 2021 year with a flying start. Clearly, the focus on our cognitive cloud networking suite is resonating with customers across diverse data sets and applications.”
How Arista Invents Products That Customers Crave
Growth comes from products that satisfy customers’ purchase criteria more effectively that competing products do. Winning companies invent new technologies that give customers much greater benefits for the money. They also track what startups are doing — to avoid missing new ideas that are being adopted by non-customers.
Arista wins because its “software-driven, data-centric approach to building their cloud architecture” better meets the needs of enterprise chief information officers for a cloud-first and data-driven network, Ullal said in the Q1 investor conference call.
Arista wins because its technology integrates “disparate functions and data sets” offered by networking incumbents — such as routers, security, switches and network management functions.
She told investors that Arista had recently won a competition for a “hospitality sector client” because its technology made it easier for the company to provision and upgrade its network while enhancing its availability and easing regulatory compliance.
Ullal is spending more of her time thinking about how to create Arista’s future. As she told me, “Most of the company is mainstream. I can add value by thinking about what we should look like in two to three years. There are many ideas and I reject many. I have three decades of networking so ideas come naturally.”
Arista finds ideas for new products from three sources. These include seeing “how the technology is evolving — [from our technology visionaries such as] Andy [Bechtolsheim, chairman and chief development officer] and Ken [Duda, chief technology officer]; Our global customer advisory board with 100 of our most intimate customers with whom we have deep friendships; and acquisitions of startups,” she said.
Arista cannot implement all these ideas. “We pick ideas that are adjacent to our business so customers can see the value; we have the ability to succeed — based on a financial forecast and business plan, and its fit with our culture. If we acquire a company that’s not a cultural fit, the merger will fail in two years,” Ullal explained.
Keys to Attracting And Motivating Top Talent
Based on my interviews with CEOs for my book, Scaling Your Startup, I learned that one key to attracting great talent — particularly in engineering — is to field a founding team of executives who are luminaries in the field. The reason is simple: ambitious people want to learn from the best.
Arista was clearly able to do this from the start. As Ullal told me, “I had a large job a Cisco. I wanted another mountain to climb. Silicon Valley is the land of entrepreneurship. Arista was different because rather than following the traditional route of taking venture capital, it was started by two technical and rich founders — Bechtolsheim and David Cheriton, a Stanford Professor, rather than taking venture capital.”
As a public company, Arista faces challenges in hiring in Silicon Valley. “Hiring talent is harder because you want to maintain the culture. Competitors in the cloud and car industries are doing well. The financial gains are not as great with a public company as they might be from joining a startup [with the potential for a successful exit],” Ullal told me.
Arista places a great emphasis on hiring people who fit its culture. As she explained, “We are customer driven and we have great people. We have a strong culture and pride. It is very hard to attract and motivate talent. You have to attract people who want to join you on a mission — not as mercenaries. We want passionate and talented people.”
One way that Arista has overcome this challenge is to broaden the geographic scope of its hiring — recruiting people from “Canada, Ireland, India, Australia, Austin, Texas, and Nashua, New Hampshire,” she said.
Empowering People to Decide And Act Quickly and Effectively
If a company encourages people closest to the customer to make decisions and act on them, it can deliver more value to customers faster than employees of companies that require multiple layers of executive approval.
On the other hand, a company that allows everyone to do their own thing will create chaos.
Arista seeks the right middle ground. As Ullal said, “We empower people who are like-minded, who understand the culture and can make decisions. We don’t hire middle managers. We want people who can do the work and who can make sure thing go right over and over again.”
This works for the majority of decisions. “90% to 95% of the decisions can be made fast, close to the customer by the systems engineer, the sales executive, and the executive who runs the region. Touch the customer, be leaders, be decision-makers. Don’t go to either extreme,” she said.
Ullal stays out of most decisions at Arista. “Many large companies lose their way. They forget what made them successful. We are proud to be public. We remain nimble and agile. 80% of the decisions are made without my involvement. I should be accessible — giving responses over the weekend if needed. We don’t want to wait for all the data to see a pattern. If I look at the data, I remember something I saw 10 years ago,” she explained.
Setting Annual Goals While Reluctantly Reporting Each Quarter
Companies that exceed analysts’ expectations and raise guidance are usually rewarded with a rising stock price.
However, that quarterly reporting game makes it more difficult to operate Arista “Setting annual goals as a public company is easier. The 90 day drill is hard. It would be better if we could report once a year. We break down the longer term — one to three years — by quarter. Long term stockholders will understand how we invest and make decisions,” Ullal said.
She does not want sales people to be so focused on quarterly quotas that they lose sight of their primary role — which is to advocate for their customers. “I do not want to hold sales people accountable for a quarterly number pulled out of the air. Sales executives should be customer advocates — not number crunching machines. They should be able to draw a blueprint of how customers can build the cloud. The best salespeople you are not sure whether they work for Arista or the customer,” she said.
Keeping Processes From Becoming Bureaucracy
Processes become more important at large companies because they ensure that the right conversations happen. However, such processes can become growth-curdling bureaucracy if they slow down or divert people from doing what is in the customer’s best interest.
Arista is questioning its processes to weed out bureaucracy. “Processes get more complicated as you go from 10 people to several thousand. People are not trying to create a bureaucracy. They follow processes because this is the way it’s done. We question old habits,” she said.
Such questioning starts at the top with Ullal wondering whether her management practices should change as the company grows. “I question my own decisions because they should be different as the company goes through different stages — private company, public, best of breed portfolio of services. Are the processes and decisions the same as in a more complex matrix?” she asked,
The processes for running an established product line are different than the ones used in trying to create a new one. “Most executives are decisive. We are the leader in the data center and the core. We are incubating new markets. For those we have to be like a startup. We have to question how well it is working and change our minds,” she said.
Over the years, Ullal has stepped away from extensive involvement in processes such as hiring and customer support. She still looks at resumes and compensation for senior level hires and used to look at customer support requests to identify patterns that might suggest product improvements.
When it comes to Arista’s decisions and processes, She is always asking “Am I adding value or adding overhead?”
Analysts Have High Hopes For Arista
Analysts have high hopes for Arista. Constellation Research Inc. analyst Holger Mueller wrote, “On the product side Arista is doing well with the focus on AI-powered security networking, and on the commercial side it grew revenue by roughly 30%, while its costs only went up 20%. Now management has to deliver in equal fashion for the rest of the year,” according to SiliconAngle.
Morningstar’s Senior Equity Analyst Mark Cash expects to see stronger revenue growth from a diverse customer base. While he thinks Arista will benefit from “Microsoft’s aggressive data center expansion plans” he sees the company’s strength in the enterprise market as shielding it from a possible spending downturn from Redmond.
Where will Arista’s new sources of revenue come from? They could be from helping customers build what Ullal calls “specialty clouds.”
As she said, “The public cloud is not right for every customer — you can go from $6,000 a month to $76 million a month as your usage increases. You need to optimize consumption — otherwise it becomes cheaper to build it yourself — which is why I like what Frank [Slootman CEO of] Snowflake does — you don’t pay for what you don’t use. There are specialty clouds — such as the WiFi cloud, the storage cloud, and application specific clouds.”