Timesdelhi.com

December 12, 2018
Category archive

national security

Unbiased algorithms can still be problematic

in Artificial Intelligence/bias/brian brackeen/Delhi/Face++/facial recognition/facial recognition software/India/kairos/law enforcement/learning/machine learning/national security/Politics/surveillance/TC/video surveillance by

Creating unbiased, accurate algorithms isn’t impossible — it’s just time consuming.

“It actually is mathematically possible,” facial recognition startup Kairos CEO Brian Brackeen told me on a panel at TechCrunch Disrupt SF.

Algorithms are sets of rules that computers follow in order to solve problems and make decisions about a particular course of action. Whether it’s the type of information we receive, the information people see about us, the jobs we get hired to do, the credit cards we get approved for, and, down the road, the driverless cars that either see us or don’t, algorithms are increasingly becoming a big part of our lives. But there is an inherent problem with algorithms that begins at the most base level and persists throughout its adaption: human bias that is baked into these machine-based decision-makers.

Creating unbiased algorithms is a matter of having enough accurate data. It’s not about just having enough “pale males” in the model, but about having enough images of people from various racial backgrounds, genders, abilities, heights, weights and so forth.

Kairos CEO Brian Brackeen

“In our world, facial recognition is all about human biases, right?” Brackeen said. “And so you think about AI, it’s learning, it’s like a child and you teach it things and then it learns more and more. What we call right down the middle, right down the fair way is ‘pale males.’ It’s very, very good. Very, very good at identifying somebody who meets that classification.”

But the further you get from pale males — adding women, people from different ethnicities, and so forth — “the harder it is for AI systems to get it right, or at least the confidence to get it right,” Brackeen said.

Still, there are cons to even a one hundred percent accurate model. On the pro side, a good facial recognition use case for a completely accurate algorithm would be in a convention center, where you use the system to quickly identity and verify people are who they say they are. That’s one type of use case Kairos, which works with corporate businesses around authentication, addresses.

“So if we’re wrong, at worst case, maybe you have to do a transfer again to your bank account,” he said. “If we’re wrong, maybe you don’t see a picture accrued during a cruise liner. But when the government is wrong about facial recognition, and someone’s life or liberty is at stake, they can be putting you in a lineup that you shouldn’t be in. They could be saying that this person is a criminal when they’re not.”

But in the case of law enforcement, no matter how accurate and unbiased these algorithms are, facial recognition software has no business in law enforcement, Brackeen said. That’s because of the potential for unlawful, excessive surveillance of citizens.

Given the government already has our passport photos and identification photos, “they could put a camera on Main Street and know every single person driving by,” Brackeen said.

And that’s a real possibility. In the last month, Brackeen said Kairos turned down a government request from Homeland Security, seeking facial recognition software for people behind moving cars.

“For us, that’s completely unacceptable,” Brackeen said.

Another issue with 100 percent perfect mathematical predictions is that it comes down to what the model is predicting, Human Rights Data Analysis Group lead statistician Kristian Lum said on the panel.

Human Rights Data Analysis Group lead statistician Kristian Lum

“Usually, the thing you’re trying to predict in a lot of these cases is something like rearrest,” Lum said. “So even if we are perfectly able to predict that, we’re still left with the problem that the human or systemic or institutional biases are generating biased arrests. And so, you still have to contextualize even your 100 percent accuracy with is the data really measuring what you think it’s measuring? Is the data itself generated by a fair process?”

HRDAG Director of Research Patrick Ball, in agreement with Lum, argued that it’s perhaps more practical to move it away from bias at the individual level and instead call it bias at the institutional or structural level. If a police department, for example, is convinced it needs to police one neighborhood more than another, it’s not as relevant if that officer is a racist individual, he said.

HRDAG Director of Research Patrick Ball

“What’s relevant is that the police department has made an institutional decision to over-police that neighborhood, thereby generating more police interactions in that neighborhood, thereby making people with that ZIP code more likely to be classified as dangerous if they are classified by risk assessment algorithms,” Ball said.

And even if the police were to have perfect information about every crime committed, in order to build a fair machine learning system, “we would need to live in a society of perfect surveillance so that there is absolute police knowledge about every single crime so that nothing is excluded,” he said. “So that there would be no bias. Let me suggest to you that that’s way worse even than a bunch of crimes going free. So maybe we should just work on reforming police practice and forget about all of the machine learning distractions because they’re really making things worse, not better.”

He added, “For fair predictions, you first need a fair criminal justice system. And we have a ways to go.”

News Source = techcrunch.com

China splits the internet while the U.S. dithers

in alibaba/Asia/Baidu/China/Cisco/Delhi/Government/huawei/India/national security/Policy/Politics/Tencent/Transsion/zte by

There are few stories as important right now as the internet being ripped asunder by the increasing animosity between the U.S. and China. Eric Schmidt, the former chairman of Alphabet, said last week at a private event in San Francisco that “I think the most likely scenario now is not a splintering, but rather a bifurcation into a Chinese-led internet and a non-Chinese internet led by America.”

He should know: Alphabet and its Google subsidiary are on the front lines of that split, experiencing a massive furor over the company’s Project Dragonfly to launch a censored search engine in the Middle Kingdom. It’s hardly alone though, with Apple facing militant criticism from Chinese netizens over its iPhone presentation and Facebook finding its application for a corporate entity on the mainland being returned and rejected.

At the heart of this split is the death of the internet as we once knew it: a unified layer for the transfer of human knowledge. As the internet has gained more and more power over society and our everyday lives, the need by governments worldwide to tame its engineering to political and moral ends has increased dramatically.

About four years ago, I wrote a piece called “From internet to internets” in which I argued that this sort of split was obvious. As I wrote at the time: “Across the world, it is becoming abundantly clear that the internet is no longer the independent and self-reliant sphere it once was, immune to the peculiarities of individual countries and their laws. Rather, the internet is firmly under the control of every government, simultaneously.”

Yet, the rules that countries like Spain put in place around media and news didn’t split the internet as I had predicted. The economic power of the U.S. and China did. Alibaba, Tencent, and Baidu may have declined in value this year, but their combined market caps is still in the trillions of dollars. WeChat, which is owned by Tencent, has more than a billion users, and while only 10% of its user base is estimated to be outside China, the ties are growing as more countries build economic bridges with the mainland.

Sometimes, those bridges are quite literal. Through the Belt and Road initiative and fledgling institutions like the Asian Infrastructure Investment Bank, China has provided massive outlays to other nations primarily around infrastructure, building partnerships and deepening economic ties.

China and the U.S. are increasingly fighting a global battle for tech legitimacy (Photo by Jason Lee / AFP / Getty Images)

That infrastructure is sometimes roads, but it can also be in areas like telecommunications. Huawei has made massive inroads into Africa, both in smartphones and in core infrastructure. Chinese-owned Transsion, which most Westerners have probably never heard of, is the dominant smartphone manufacturer on the continent.

Chinese-made telecom infrastructure. Chinese handsets. Increasingly Chinese apps. For all of the concerns of Congress and national security officials about Huawei and ZTE equipment entering the American or Australian markets, the real fight for the future of the internet is going to be in precisely these developing regions which have no incumbent technology.

That’s what has made the Trump administration’s strategy toward trade negotiations with China so miserable to watch. The focus has been on repeated rounds of tariffs that will ensure that Chinese goods — particularly in high-tech industries — are more expensive to American consumers, allowing domestic manufacturers to better compete. Yet, the policies have done nothing to ensure that American values around the internet are exported to continents like Africa or South America, or that Cisco’s equipment will be chosen over Huawei’s.

That might be changing at long last. The Financial Times reported yesterday that the Trump administration is preparing to double down on the Overseas Private Investment Corporation, which offers commercial lending facilities to developing countries. It would be merged into another agency and given a much more rich budget (as high as $60 billion) to go and compete with Chinese financing around the world.

Maybe that measure will be successful in closing the strategic distance between the two countries. Maybe rumors that the administration is going to broadly double down on the trade war will lead to a much more comprehensive set of policies.

But along the way, regardless of what happens, these skirmishes will lead to a fracturing of the internet, and along with it, the death of the internet as a bastion and voice of freedom and knowledge for all people everywhere.

News Source = techcrunch.com

Russian indictments show that the U.S. needs federal oversight of election security

in America/Column/Congress/cybercrime/defcon/Delhi/democratic party/Department of Homeland Security/Election Assistance Commission/election security/elections/federal election/Federal government/Florida/Government/helsinki/India/national security/operating systems/Politics/president/presidential election/Ron Wyden/Russia/Trump/United States by

President Trump’s Helsinki summit with Vladimir Putin, on the heels of twelve Russian intelligence officials indicted for hacking the 2016 election, made it clear that this administration has zero commitment to protect our elections from future Russian attacks.

These events should remind us of an alarming fact we can no longer afford to ignore: our elections are not secure.

As a nation, we underfund and neglect election security. So, much like our aging infrastructure, our election infrastructure is severely outdated and crumbling before our eyes.

Unfortunately, in today’s hyper-partisan environment, even concerns over election security are divided along party lines. Case in point: after his trip to Russia last week, Republican Senator Ron Johnson declared “It’s very difficult to really meddle in our elections. It just is.”

To effectively safeguard our elections, we need to consider yet another conservative taboo: the federal government should have more power in setting election security standards. Our current decentralized, disjointed state-based system is no longer adequate for protecting our elections against foreign interference in the 21st century.

TechCrunch/Bryce Durbin

Right now, the federal government plays a very limited role in the oversight of election security. The Election Assistance Commission and Department of Homeland Security offer optional resources and issue non-binding guidelines for best practices, and states are free to come up with their own standards as they please. The results, unsurprisingly, are abysmal.

In 2016, for example, over two-thirds of all counties in the U.S. used voting machines that were over a decade old. Many machine used outdated softwares and ran in absurdly old operating systems such as Windows 2000. Thirteen states still use machines that are completely electronic, which makes themprone to glitches, and with no paper trails, the results cannot be audited.

Many experts have pointed out that our current machines could be hacked in a matter of minutes. Recently, a 14 year-old participant at DefCon breached a voting machine in 90 minutes, and was able to change the vote tally in the machine remotely, from anywhere.

Besides the machines, there are other major vulnerabilities in many states’ election security standards that would make hacking our elections a breeze for the Russians. Our voter registration databases are outdated and prone to infiltration. Many states have no post-election auditing requirements at all, and those that do are often insufficient, severely undermining our ability to identify and correct an attack.

While federalizing election security has long been castigated as an infringement of state rights, politicians are beginning to acknowledge its necessity. Senator Ron Wyden, for instance, recently introduced The Protecting American Votes and Elections Act of 2018, whichwould require every state to use election machines with paper ballots and mandate risk-limiting post-election audits (the “gold standard” of election auditing).

As Wyden argues: “Americans don’t expect states, much less county officials, to fight America’s wars. The Russians have attacked our election infrastructure and leaving our defenses to states and local entities, in my view, is not an adequate response. Our country needs baseline, mandatory, federal election security standards.”

TechCrunch/Bryce Durbin

Rather than providing concrete solutions, this Republican Congress continues to pretend that all of our election security problems can be solved by tiny, poorly designed federal grant programs alone. In this year’s omnibus spending bill, a bipartisan compromise provided a meager, but much needed $380 million federal grant to states for strengthening election security ahead of the 2018 election. However, the effectiveness of this grant is questionable, given it was earmarked for broad purposes and allocated by a formula that is not competitive or need-based.

Worse still, since states are not required to spend the federal grant allocated to them, some stateshave not even applied to collect their shares. Several state governments are impeding the use of this grant through a combination of delayed action and inaction. For example, Florida’s Republican-led state legislature has refused to authorize their election officials to use the grant before the 2018 election, even when the state is in desperate need for more election security funding.

While inadequate funding is a serious concern that needs to be addressed — House Democrats estimated that we will need $1.4 billion over the next decade to bring our entire election system in line with best practices — increasing federal grants alone would not be enough to secure elections in every state. The Secure Elections Act, a bill currently with the most broad-based, bipartisan support, will provide much needed federal funding to make up for the current shortfall, but as with this year’s federal grant, there is no guarantee states would use the funding in a timely and effective fashion — or at all — given state participation will remain voluntary under this bill.

Our representative democracy cannot survive if we fail to preserve the fairness and integrity of our elections. While it’s too late to implement binding federal guidelines to secure the 2018 midterm, we should accept nothing less for the 2020 presidential election, as we can be certain the Russians will hack that election in order to help their preferred candidate, yet again.

Too many states have proven they are unwilling to take election security seriously. It’s time for the federal government to step in.

News Source = techcrunch.com

The United States needs a Department of Cybersecurity

in China/Column/computer security/Congress/cyberattack/cybercrime/Cyberwarfare/Delhi/department of defense/Department of Homeland Security/department of justice/executive/Federal Bureau of Investigation/Government/hacking/India/national security/Politics/Russia/San Francisco/Security/spy/United States/Washington by

This week over 40,000 security professionals will attend RSA in San Francisco to see the latest cyber technologies on display and discuss key issues. No topic will be higher on the agenda than the Russian sponsored hack of the American 2016 election with debate about why the country has done so little to respond and what measures should be taken to deter future attempts at subverting our democracy.

For good reason. There is now clear evidence of Russian interference in the election with Special Counsel Mueller’s 37-page indictment of 13 Russians yet the attack on US sovereignty and stability has gone largely unanswered.  The $120 million set aside by Congress to address the Russian attacks remains unspent. We expelled Russian diplomats but only under international pressure after the poisoning of a former Russian spy and his daughter.

Recent sanctions are unlikely to change the behavior of the Putin administration. To put it bluntly, we have done nothing of substance to address our vulnerability to foreign cyberattacks. Meanwhile, our enemies gain in technological capability, sophistication and impact.

Along with the Russians, the Chinese, North Koreans, Iranians and newly derived nation states use cyber techniques on a daily basis to further their efforts to gain advantage on the geopolitical stage. It is a conscious decision by these governments that a proactive cyber program advances their goals while limiting the United States.

Krisztian Bocsi/Bloomberg via Getty Images

We were once dominant in this realm both technically and with our knowledge and skillsets. That playing field has been leveled and we sit idly by without the will or focus to try and regain the advantage. This is unacceptable, untenable and will ultimately lead to potentially dire consequences.

In March of this year, the US CyberCommand released  a vision paper called “Achieve and Maintain Cyberspace Superiority.” It is a call to action to unleash the country’s cyber warriors to fight  for our national security in concert with all other diplomatic and economic powers available to the United States.

It’s a start but a vision statement is not enough.  Without a proper organizational structure, the United States will never achieve operational excellence in its cyber endeavors.  Today we are organized to fail.  Our capabilities are distributed across so many different parts of the government that they are overwhelmed with bureaucracy, inefficiency and dilution of talent.

The Department of Homeland Security is responsible for national protection including prevention, mitigation and recovery from cyber attacks. The FBI, under the umbrella of the Department of Justice,  has lead responsibility for investigation and enforcement. The Department of Defense, including US CyberCommand, is in charge of national defense.  In addition, each of the various military branches  have their own cyber units. No one who wanted to win would organize a critical  capability in such a distributed and disbursed manner.

How could our law makers know what policy to pass? How do we recruit and train the best of the best in an organization, when it might just be a rotation through a military branch? How can we instantly share knowledge that benefits all when these groups don’t even talk to one another? Our current approach does not and cannot work.

Image courtesy of Colin Anderson

What is needed is a sixteenth branch of the Executive — a Department of Cybersecurity — that  would assemble the country’s best talent and resources to operate under a single umbrella and a single coherent policy.  By uniting our cyber efforts we would make the best use of limited resources and ensure seamless communications across all elements dealing in cyberspace. The department would  act on behalf of the government and the private sector to protect against cyberthreats and, when needed, go on offense.

As with physical defense, sometimes that means diplomacy or sanctions, and sometimes it means executing missions to cripple an enemy’s cyber-operations. We  have the technological capabilities, we have the talent, we know what to do but unless all of this firepower is unified and aimed at the enemy we might as well have nothing.

When a Department of Cybersecurity is discussed in Washington, it is usually rejected because of the number of agencies and departments affected. This is code for loss of budget and personnel. We must rise above turf battles if we are to have a shot at waging an effective cyber war. There are some who have raised concerns about coordination on offensive actions but they can be addressed by a clear chain of command with the Defense Department to avoid the potential of a larger conflict.

We must also not be thrown by comparisons to the Department of Homeland Security and conclude a Cybersecurity department would face the same challenges. DHS was 22 different agencies thrust into one. A Department of Cybersecurity would be built around a common set of skills, people and know-how all working on a common issue and goal. Very different.

Strengthening our cyberdefense is as vital as having a powerful standing army to defend ourselves and our allies. Russia, China and others have invested in their cyberwar capabilities to exploit our systems almost at will.

Counterpunching those efforts requires our own national mandate executed with Cabinet level authority. If we don’t bestow this level of importance to the fight and set ourselves up to win, interference in US elections will not only be repeated …  such acts will seem trivial in comparison to what could and is likely to happen.

News Source = techcrunch.com

Lessons from cybersecurity exits

in Adobe/Bain Capital/Business/ceo/chief information security officer/CIO/Column/computer security/CTO/cyber/cybercrime/Cyberwarfare/Delhi/Director/Economy/Entrepreneurship/head/India/IPOs/LinkedIn/Marc Andreessen/mulesoft/national security/nea/Okta/palo alto networks/partner/Phantom Cyber/PhishMe/Politics/Private equity/sailpoint/Sendgrid/splunk/Symantec/TPG Growth/trident capital/True Ventures/Twilio/unicorn/venrock/Venture Capital/zscaler by

To: ceo@cybersecuritystartup.com

Subject: Lessons from cybersecurity exits

Dear F0und3r:

What a month this has been for cybersecurity! One unicorn IPO and two nice acquisitions – Zscaler’s great debut on wall street,  a $300 million acquisition of Evident.io by Palo Alto Networks and a $350 million acquisition of Phantom Cyber by Splunk has gotten all of us excited.

Word on the street is that in each of those exits, the founders took home ~30% to 40% of the proceeds. Which is not bad for ~ 4 /5 years of work. They can finally afford to buy two bedroom homes in Silicon Valley.

Evident.IO Investment Rounds and Return estimates

Date

Select Investors

Round Size

Pre

Post

Dilution

Estimated Returns / Multiple of Invested Capital

Sep 2013

True Ventures

$1.5m

$5.25m

$6.75 m

22%

44X

Nov 2014

Bain Capital

$9.8 m

$18.1m

$28.0 m

35%

10.7X

Apr 2016

Venrock

$15.7 m

$35.0 m

$50.7 m

30%

6X

Feb 2017

GV

$22.0 m

$73.6 m

$95.5

23%

3.1X

My math is not that good but looks like even some VCs made a decent return. Back of the envelope scribbles indicate that True Ventures scored an estimated ~44X multiple on its seed investment. Others like Bain snagged a ~10X on the A round investment and Venrock which led the Series B round took home ~6X.

We see a similar pattern with Phantom Cyber, which got acquired by Splunk for $350 million. A little bird told me that they had booking in the range of $10 million. But before we all get too self-congratulatory, lets ask – why did these companies sell at $300 million to $350 million when everyone in the valley wants to ride a unicorn? Clearly, funds like GV, Bain and Kleiner could have fueled more rounds to make unicorns out of Evident.io and Phantom Cyber.

Phantom Cyber Investment Rounds and Return estimates

Date

Select Investors

Round Size

Pre

Post

Dilution

Estimated Returns / Multiple of Invested Capital

April 2015

Foundation Capital

$2.7m

$8.3 m

$11.04 m

14.50%

31.7

Sep 2015

Blackstone

$6.5m

$26.7 m

$33.2 m

15.90%

10.5

Jan 2017

KPCB

$13.5m

$83.0 m

$96.5 m

13.90%

3.6

(Data Source: Pitchbook)

Some of the board members might have peeked at the exit data gathered by the hardworking analysts at Momentum Cyber, a cybersecurity advisory firm. Look at security exit trends from 2010-2017. You might notice that ~68% of security exits were below $100 million. And as much as 85% of exits occur below $300 million.

Agreed that there are very few exceptional security CEO’s like Jay Chaudhry who grew up in a Himalayan village, and led ZScaler to an IPO. This was Jay’s fifth startup and he kept over 25.5% of the proceeds, with another 28.3% owned by his trust. TPG Growth owned less than 10%. After all, he himself funded a substantial part of the company (which raised a total of $110 million).  But not everyone is as driven, successful and it’s ok to sell if the exit numbers are meaningful. Remember what that bard of avon once said:

For I must tell you friendly in your ear,

Sell when you can; you are not for all markets.

(Shakespeare, As you Like It, Act 3, Scene V)

(68% of security exits occur below $100 million. M & A Data from 2010-2017. Source: Momentum Cyber)

My friend Dino Boukouris, a director at Momentum Cyber, offers some sage advice to all founders who are smitten by unicorns. “Before a founder raises their next round, I would reflect on the market’s ability to purchase companies. The exit data says it all. As you raise more capital, your exit value goes up, timing gets stretched and the number of buyers who can afford you drops.” Dino has a point, you see. As we inflate valuations, your work, my dear CEO, becomes much harder.

If you don’t believe Dino, let’s look at another recent exit, PhishMe, which was acquired by a private equity consortium for $400 million. That’s a nice number, correct? At the first look, you’ll notice that the dilution and financial return patterns are similar to that of Phantom. Except that PhishMe took 7 years and consumed $58 million of capital, while Phantom took 3 years and consumed $22.7 million. Timing and capital efficiency matter as much as exit value. It’s not just the exit value ~ but how long and how much. Back to my man, Dino who will gently remind you that for the 175 M & A transactions in 2017, the median value was $68 milion. Read that last sentence again — very slowly. $68 million. Ouch!

PhishMe Investment Rounds

Date

Round size

Select Investors

Pre-money Valuation

Post

Dilution

Returns / Multiple of Invested Capital

July 2012

$2.5m

Paladin

$10 m

12.5 m

12.20%

32.0

March 2015

$13 m

Paladin

$61 m

$74 m

13 %

5.4

July 2016

$42.5 m

Bessemer

$155 m

197 m

21%

2.0

(Data Source: Pitchbook)

Two years ago  in Cockroaches versus Unicorns – The Golden Age of Cybersecurity Startups cybersecurity founders were urged to avoid the unicorn hubris. A lot of bystanders, your ego included, will cheer you as you get higher valuations. But aren’t we all rational human beings, always making data based decisions?

Marc Andreessen will remind you that his best friend, Jim Barksdale, once said “If we have data, let’s look at data. If all we have are opinions, let’s go with mine.”   Since 2012, my VC friends have funded 1242 cybersecurity companies, investing a whopping $17.8bn. But chief information security officers say that they don’t need 1242 security products. One exhausted CISO told me they get fifteen to seventeen cold calls a day. They hide away from LinkedIn, being bombarded relentlessly.

Enrique Salem (former CEO of Symantec) and Noah Carr, both with Bain Capital are celebrating the successful sale of Evident.io. They pointed out that the founders — Tim Prendergast and Justin Lundy had lived the public cloud security problem in their previous lives at Adobe. “Such deep domain expertise allowed them to gain credibility in the market. It’s not easy to earn the trust of their customers. But given their strong engineering team, they were able to build an “easy to deploy” solution that could scale to customers with 1000s of AWS / Azure accounts. Customers were more willing to be reference-able, given this aligned relationship.”

(Source: Momentum Cyber)

You, my dear CEO, should take a page from that playbook. Because Jake Flomenberg, Partner at Accel Partners says, “CISOs are suffering from indigestion. They are looking to rationalize toolsets and add very selectively. New layer X for new threat vector Y is an increasingly tough sell.” According to Cack Wilhelm Partner at Accomplice, “Security analysts have alert fatigue, and CISOs have vendor fatigue.”  You are one of those possibly, wouldn’t you agree?

Besides indigestion and fatigue, the CISO roles have become demanding. William Lin, Principal at Trident Capital Cyber, a $300m fund pointed out that “the role of CISO has bifurcated into managing risk akin to an auditor and at the same time, managing complex engineering and technology environments.”  So naturally, they are managing their time more cautiously and not looking forward to meeting one more startup.

Erik Bloch, Director of Security Products at SalesForce says that while he keeps an open mind and is willing to look at innovative startups, it takes him weeks to arrange calls with the right people, and months to scope a POC. And let’s not forget the mountain of paperworks and legal agreements. “It’s great to say you have a Fortune 100 as an early customer, but just be warned that it’ll be a long, hard road to get there, so plan appropriately” he pointed out.

So, my dear founder, as the road gets harder, funding slows down. Look at 2017 —  despite all those big hacks, Series A funding dropped by 25% in 2017. Clearly, many of our seed funded companies are not delivering those Fortune 100 POC milestones. And are unable to raise a Series A.  Weep, if we must, but let us remind ourselves that out point solutions are not that impressive to the CISOs.

All the founders I know are trying to raise a formulaic $8m Series A on $40m pre. But not every startup that wants 8 on 40 deserves it. Revenues and growth rate, those quaint metrics matter more than ever. And some investors look for the quality of your customers.  Aaron Jacobson of NEA, a multi-billion dollar venture fund says, ”A key value driver is a thought-leader CISO as a customer. This is often a good indicator of value creation.“

Stage

Expected Revenue Run Rate

Estd. Round Size

Angel

None

Up to $2m

Series A

$1.5m to $3 m

$5m to $8m

Early VC

$5 m to $8 m

$15m to $25m

Late Stage VC

$6m to $10m

$30m to $50m

When markets get crowded and all startups sound the same, investors seek quality, or move to later stages.  They like to see well proven companies, that have solved a lot of basic problems. And eliminated riskier stumbling blocks. Like product-market fit, pricing and go-to-market issues. Naturally, the later stage valuations are rising faster. Money is chasing quality, growth and returns.

Median Post-Money Valuation by stage for cybersecurity companies (Source: Pitchbook)

The security IPOs offer a sobering view. This is a long journey, not for the faint of heart. Okta moved fast, consumed ~4X more capital as compared to Sailpoint and delivered great returns.

Company

Year Founded

Years to IPO

Total Capital raised prior to IPO

Revenues (2017)

Post Money of last round prior to IPO

Market Cap at IPO

ZScaler

2008

10

$180m

$176 m

$1.05 bn

$3.6 bn

Okta

2009

8

$231 m

$160 m

$1.18 bn

$2.1 bn

Forescout

2000

17

$159 m

$220 m

$1.0 bn

$806 mn

SailPoint

2004

13

$54.7 m

$186 m

N/A

$1.1 bn

Security IPOs (Source: Momentum Cyber, Pitchbook)

Innovating with go-to-market strategies

In the near term,  the big challenge for you, dear security founder, is selling in an over crowded market. If I were you, I’d remember that innovation should not be restricted to merely technology, but can extend into sales and marketing. We lack creativity when it comes to marketing – ask Kelly Shortridge of Security ScoreCard. She should get some kind of BlackHat award for developing this godforsaken Infosec Startup Bingo. If you find any startup vendor that uses all these words, and wins this bingo, please DM me ~ I will promptly shave my head in shame. We got here because we do not possess simple marketing muscles. We copy each other while our customers roll their eyes when we pitch them.

Sid Trivedi of Omidyar Technology Ventures wants to work with the developer focussed startups. He says, “Look at companies like Auth0. The sales efficiency on developer-focused platforms is tremendous. You can go to a CISO, CIO or CTO and point out that X number of developers are paying to use my technology. Here are their names, why don’t you talk to them? And then, let’s discuss an enterprise license for the full company?” That approach works like magic. Overwhelming majority of the software IPOs like Twilio, Mulesoft, SendGrid are developer platforms.”

If you go top-down in a hurry, you can crash and burn. I am aware of an impatient security vendor who used executive level pressure at a Fortune 50 company. They kicked their way into the POC. And got kicked out by the infosec team. The furios infosec team destroyed the vendor in a technical assessment. I was told that the product was functional but the vendor’s impatience and political gymnastics killed the deal. Let us not forget simple truth: many times CISOs turn to their subordinates for advice and decision-making, so don’t just sell to the top. Nor ignore the rest of the people in the room.

With more noise, the buyers freeze. Margins shrink. Revenues and growth slows down. Which means it’s harder to get to your milestones before your next round. Running out of cash is not fun. Nor is a down round, layoffs and such. So while this is easier said than done, please raise less and do more. And maybe, just maybe, you can keep 40% of a $350 million exit.

If you have questions or existential dilemmas, you can always find me, chatting with a friendly VC in South Park.  Or I’m always around in a trusted secure world of Signal.

Stay safe at that annual security stampede called RSA.

Kindly,

Mahendra

PS: Let’s not forget to express our gratitude to those analysts at Momentum Cyber and Pitchbook for painstakingly tracking every investment, analyzing and presenting meaningful data. They help us look at the forest, and make our journey easier. Send them a thank-you tweet, some wine, chocolates, flowers or home-baked cookies.

News Source = techcrunch.com

1 2 3
Go to Top