What I Learned Working At A “Scale-Down”
Scaling a company is an amazing learning experience, but what about watching a company fall apart?
I was part of the team that built AnyConnect when I worked in Cisco’s Security Business Unit. I still remember when a Product Manager and I pitched the idea for building AnyConnect, and many thousands of code lines later, it had a footprint of 100 million seats. (I’ll admit it, I’m really proud of that.)
As smartphones are way more useful than feature phones, so VPN was lightyears ahead of what preceded it — a thing known as remote access. The way you got into your enterprise network was over modems. These bad boys were not quite as ancient as, say, the wheel or fire, but they were primitive.
Bandwidth was measured in Kbps (kilobits per second), not Mbps (megabytes per second). The Boomers here are probably groaning in remembrance. These things were brittle as corn flakes and dog slow. Think of it this way: if you wanted to download, say, Moonlight, it might take a week and a half. That slow.
The early front-runner in remote access was Shiva Corp. I worked there as a software engineer, and later a software manager. We built a box that sucked in modem data and translated data streams on the enterprise net to IP, IPX, and Appletalk. I wrote a lot of Appletalk code.
We sold a crap-ton of those boxes.
Shiva went public with a bang: the stock doubled on the first day of training. And doubled again. And doubled again! We were scaling like mad — hiring engineers, building market share, refining our sales motion, writing a boatload of new software, just growing like a weed.
Three years later, Shiva Corp. was sold to Intel for less than the IPO market cap. As the co-founder had predicted to me years earlier “We are either going to make a huge dent or leave a big crater.” Well, we definitely left a huge honkin’ crater.
Hundred of millions of investor dollars wiped out in about a year, sold to Intel for a pittance. What the heck happened?
At Shiva, here’s what we fumbled when scaling up, resulting in our slide down:
1. Tech scale-ups must hit critical inflection points.
This is deliberately #1; if you take away nothing else, please take this one away. If life is change, tech is life on steroids. There are critical inflection points. You could also call them paradigm changes.
Mainframe to mini-computer.
Mini-computer to laptop.
Feature phone to smartphone.
Enterprise data center to public cloud.
If a paradigm change impacts your market, your customers, and you miss it, you are probably doomed. Behemoths, like Microsoft or Intel, can buy their way out of a miss. Scale-ups might, but it’s a risky strategy, especially early in the scale-up. You have to anticipate these inflection points and invest as needed.
2. If/when you hit the point where you are replacing a founder CEO, don’t f*ck it up.
Our co-founder was inexperienced as a CEO, and the co-founder and board both felt that we needed someone more experienced to scale up. That’s good. The Board hired someone who was visionary and smart and talked a good game, but made incredibly bad decisions. That’s bad. Really bad. You cannot afford a miss on this decision during the scale-up stage.
3. Don’t play favorites.
The engineering culture was rife with favoritism. There was an “in” group. You can’t scale up a strong team based on that culture. That doesn’t mean you peanut butter rewards. It does mean you treat everyone with respect, and promote inclusion. The cool “in” gang that was there early on failed to resist the temptation to have an attitude. No-one coached them to do better (see #6 and #15).
4. Take your lumps; don’t mortgage your future.
Shiva leveraged a channel-based sales model. Revenue was recognized when we sold boxes into the channel. This revenue recognition model called for tremendous discipline. We didn’t have it.
As soon as hitting a quarter’s numbers looked risky, Shiva stuffed the channel. Boxes sat in channel partner storerooms. This is a drug habit. It only gets worse over time. At it’s inevitable — one day, the the bill comes due.
5. Build partnerships when you are strong, not to save your butt.
After we hit friction, we turned to a partnership with Nortel to leverage their behemoth sales force. When you’re down, you lack leverage, and you sign bad deals.
We did.
Also, if you’re going to hitch your wagon to someone else’s team, make sure the horses are healthy. (PS: Nortel went bankrupt.)
6. When you hire execs, hire for emotional intelligence.
We hired an Engineering VP that had zero people skills. Motivation plummeted.
7. Don’t ship hype.
Our cash-cow was the “LanRover.” When the web got hot, we bundled a LanRover with a separate wide area connector box and called it a “WebRover.” When we announced it, the stock skyrocketed. We maybe sold a hundred? All hype, no customer value.
8. If you need to invest, don’t play at it.
We eventually saw the hand-writing on the wall, and realized we needed VPN. A few folks were identified to build a VPN solution, and the one they built (I was gone by then) was a day late and a dollar short. Shiva should have bought a top-notch product, or invested heavily to build something serious.
9. Tech companies need to build interesting things.
The best engineers like to build things. They will fix bugs and refactor, but will tolerate only so many variations on a theme. At some point, eventually, they need to build something they think is cool. When they haven’t for a long time, and see nothing on the horizon, they leave. And they did.
10. Think long term.
Shiva fell into the “think in quarters” trap that plague some companies after they IPO. Yes, you have to keep your investors happy. That alone is not enough. At Shiva, we failed to build on a dream.
11. Culture eats strategy for breakfast.
We didn’t have a Chief of People Ops or any kind of culture team invested in guiding our culture. As a result, the air grew toxic. It was avoidable.
12. But have a strategy.
If we had one, I don’t know what it was. Hitting your quarterly numbers to keep your investors happy isn’t a strategy.
13. Design matters.
We had really good engineers. We had no designers. Our engineers had no idea how to create a delightful user experience.
14. Buy wisely and integrate thoughtfully.
We bought two companies. One of them made software to optimize IE performance. History is littered with defunct software companies that tried to capitalize on Windows limitations. This was not hard to forecast.
The other company was in Scotland. There was little effort to integrate the two engineering teams. So, we had two separate engineering teams. You can imagine the results.
15. Train managers.
We didn’t. We had engineers who became managers who knew nothing about managing or leading, same with sales and marketing. I became a manager. I wasn’t very good, I had no training.
Learning to be a good manager is really freakin’ hard. Decades later, I am still learning. It doesn’t just happen by magic. If your managers aren’t good, you are just as doomed as if your individual talent isn’t good, because the good individual contributors won’t hang around.
16. Have an active CTO.
This is related to the first point. Someone needs to own this, even if it’s not their title. We had no-one doing this. Tech companies need people who see ahead, who are directly responsible for seeing ahead.
17. Hire “and” superstars.
Our best engineers, and some of them were really quite good, had a boatload of attitude. We didn’t have anyone like our Senior Tech Leads at HubSpot, who are stellar technically and also great people leaders, who treat people of all abilities with respect. You must have both in at least some of your superstars.
18. Diversify but thoughtfully.
At Shiva, we had remote access and only that. The earlier AppleTalk products died when IP took over the world. We should have moved into adjacencies. (Within reason, of course: when I was at Cisco, the company diluted its strength by developing dozens of adjacencies.)
19. Lean into failure.
We never talked about it when we screwed up. Executives never showed vulnerability. This was an epic fail.
20. Ignore the stock price.
It became a thing. Who had options, who didn’t, what was the stock price going to do. The stock price was an inevitable consequence of strategy and execution; paying attention to it only served to distract us from the mission.
