Friday, May 1, 2020

Startup Offense and Defense in the Recession

Multiple economic indicators suggests the economy is getting worse. Like Wiley Coyote suspended in mid-air and not yet realizing he is about to plummet to the ground, the US economy is magically suspended by wishful thinking and unsustainable monetary policies.

Emily Cunningham Twitter वर: "Oil demand peaking in 2020s means ...
^Current image of the US Economy (source: US Fed)

A number of economic measures to give a sense of the situation.

Unemployment claims at all time highs:

Retail sales down dramatically.

Factory output down to lowest levels since post-WW2.

Restaurants in a mass shut down. OpenTable data for restaurants.

Hospital systems are losing large amounts of money due to a lack of elective procedures.
Revenue at many non-profit hospitals has fallen 50% or more in the last few weeks.

Cities and municipalities are heading towards large budget deficits due to drop in tax receipts.

China data does not lend much comfort to a fast recovery. Production is down 10% and consumer demand is down 30%, despite China "unlocking". China is still grinding back to baseline and has a long way to go. We should expect the same of the US economy.

Economic shocks and cascades
It usually takes 3 to 6 months for an economic shock to wend its way through the economy. Like dominos slowing falling over there is a cascade of hiring freezes, purchasing freezes and layoffs. The companies whose customers did purchasing freezes or cost cuts suddenly find their own revenue dropping, so they too need to decrease costs in headcount and/or purchasing.

Right now big enterprises have been going through the following steps:
1. Figure out this COVID thing. (1-2 weeks)
2. Figure out how to get 10,000 people to work from home and become somewhat productive. (4 weeks)
3. Figure out impact of economic slowdown to our business and customers (2-4 weeks)
4. Decide how to cut costs and pull together layoff plan (4 weeks)
5. Take action like layoffs, IT buying freeze etc. (4-8 weeks)

In other words, it should take a large enterprise 3-6 months to move to remote work, understand its financial situation, plan a layoff and buying freeze, and execute it. Given that we are only a month or two into COVID, much of the damage is yet to come. This means many startups won't see the real impact of the slow down for anywhere between a few weeks and two quarters as big enterprises figure it all out and then act.

As more people become unemployed (many economists estimate 20-30% unemployment) and consumer demand drops, more large enterprises will take the following actions:
1. Slow or freeze purchasing
2. Renegotiate prices or contracts & backend payments
3. Freeze hiring or do layoffs
4. Divest assets or divisions
5. Some will also be aggressive and start doing more M&A or market consolidation

As a startup you will need to make a series of offensive and defensive moves to deal with this new environment.

If you are an early stage company that has raised 2-3 years of cash, you may be fine and can ignore a lot of the advice below. Your focus should be on building a team and product that will get to product/market fit in today's environment. The main suggestion would be to focus on a product that can sell in a year when your product is ready - the biggest risk for an early stage company is that sales cycles may take longer.

For everyone else, read on :)

Defensive Moves
1. Check your cash burn. 
Assess your situation:
How much cash do you have?
What is happening to your business?
What is likely to happen given customer mix?

You need to figure out how to get to 2 to 3 years of cash. 3 is optimal as it gives you:

  • 6-12 months for the economy to bottom
  • 6-12 months to adjust and start to grow again. You will need to show growth to fundraise at a good price.
  • 9-12 months of buffer when you go out to fundraise

Getting to 2-3 years may include fundraising, pricing or payment changes with customers, or cutting your own costs via layoffs or other means.

In planning burn you should assume also do a "stress test" model. For example, if you planned to grow 3X this year, what happens if you grow 2X? If you are flat with the year? If you are down 50%? Some businesses are actually accelerating out of COVID and this might not be relevant. If you are already seeing things slow you might want to see what a low probability, worst case plan means for cash management as the economic picture unfolds during the next 2 quarters.

2. Raising money in a recession.
There is an old saying in fundraising "when they are passing the hors d'oeuvres take two" - when money is available you should take it. Fundraising in a recession is different from fundraising in good times.

  • Optimize for an OK valuation and speed versus a great valuation and slow process. Valuations that look cheap in a given moment may look expensive 6 months later during volatile times. In good times you should optimize for valuation. In bad times, you should optimize for speed, simplicity, and buffer.
  • Take a bit of extra money. Make sure that in your very worst case you can have 3 years of cash or bridge to profitability. A little extra dilution in a recession is better than uncertainty on multiples or cash position in the future.
  • Remember that getting profitable gives you more options than not. The lower your burn, the higher your leverage all else being equal.
  • Expect more inside rounds. Many VCs are focusing first on their own portfolio in terms of funding, and then will look externally in a few months. Expect many of the rounds in the next few months to be driven by insiders. In some cases these will be reopenings of prior rounds.
  • Pricing uncertainty. VCs are struggling with how bad businesses will be impact due to the recession. Given that the next 3-6 months are highly uncertain VCs do not know what valuation to give. Will the business still grow its projected 300% this year or will it growth 50%? Enterprise sales deals from last year are closing in April, but will new sales deals initiated remotely last month close or stay open longer? Given the business uncertainty many VCs are having trouble giving offers, making fundraising harder.
  • May and June are likely to be months of relative optimism. The stock markets are resilient. Lockups are coming off. The sun is out. Big promises are being made on vaccines and drugs (without data to support them yet). The economic carnage has not been highlighted by media much yet. The next 2 months may be a relatively good time to fundraise.

3. Understand which customers are at risk, and which should be embraced.
If you sell to enterprises, go through your customer list. If you have revenue concentration, look at your top N accounts. Do you have travel, local governments, retail, real estate, or other customers? If so, which are high risk of cancelation or renegotiation? You can discount this revenue as you plan for burn.  You may need to fire bad customers, or stop extending credit.

Which customers also seem to be accelerating or doing better in this time? You may want to spend more time with them or focus your sales efforts on upselling your most stable customers or getting your product prepaid.

If you do not have high revenue concentration, you can instead look at vertical views. What % of your business is exposed to travel, DTC, offline retail, restaurants or other segments? This will allow you to predict customer churn or models and understand runway. You may also want to cut marketing or inside sales to verticals that are likely to churn and reallocate spend to verticals that will be stable or grow.

4. Plan for hiring freezes or layoffs.
Although it takes time to build a recruiting pipeline, when in doubt you may want to freeze hiring and wait to see the impact the recession will have on your business. If you think there is some chance you will need to do a layoff, plan it in advance. It often takes some time to execute and you want to do it well, both for the people leaving as well as the employees who stay behind.

The key tenets of a layoff include:

  • Have a crisp plan and process.
  • Treat people well.
  • Create stability for the people who are still with the company. This layoff's goal is to save their jobs and save the company, not remove other people's jobs.
  • Cut deeper than you ever think you will need to. It is much worse to do multiple sequential layoffs than a single deep cut. You can always rehire people if you cut too much, but it is tough and hurts the culture to do multiple rounds.

When I moved out to Silicon Valley I joined a high flying startup backed by Sequoia and Matrix. The company grew from 120 to 150 people, and then shrank to ~12 people over 5 rounds of lay offs as the internet bubble imploded. I got laid off in the 3rd round. It would have been better for the company to do one big cut up front and preserve cash than to wait through multiple layoffs and keep burning large sums of money. Repeated layoffs devastated moral at the company.

A16Z has a good guide to planning layoffs here.

5. Build culture and create stability for employees
While many founders thrive on unpredictability, most employees crave stability. Particularly in times like these it is important for people to feel safe and secure. Employee stability may include create and communicating the following:

  • Financial stability of the company. Explain cash position, customers, and how the company is going to bridge the COVID-19 recession.
  • COVID plans and policies. People have uncertainty about whether they can go back into the office and under what conditions to do so. You can put together a plan around office reopening. Some companies have told employees that they will have the *option* to work from home through e.g. September, so employees can plan out a few months of their life in advance. Each state has its own requirements for businesses to reopen so this can get complicated if you are a larger company or deal with manufacturing.
  • Build ties. Bonds help in times of adversity. Can you do a zoom-based speaker series? Drinks for the team? Think of ways to get people together or bonding in small groups.
Offensive moves
There is an old saying to "never waste a good recession". While recessions can be daunting, scary times it may also create opportunities for your business.

1. Reposition your company and grow faster. A number of companies are actually doing better due to COVID. If you are working on online collaboration, edtech, or certain vertical SaaS, you may be doing better than ever. You may also want to reposition your company's marketing materials and narrative towards cost-cutting, remote collaboration, outsourcing, or other topical areas.

Zoom, Instacart, and others probably gained multiple years of customer adoption and enterprise growth due to this new environment.

2. Hire great people. Great people are coming onto the market due to layoffs, company shut downs (should see more of those in 6-12 months) and just general feelings of instability. If you are a stable, growing company now is a great time to target specific amazing people to hire.

3. M&A time! In the next year or two there will be a lot of opportunities to either buy other companies and grow into new areas, or to exit. Some companies that are seeing fast, non-sustainable growth due to work-from-home and social distancing may want to exit while things look good (looking at you much of EdTech!).

4. Build your moat, build your infrastructure. The next few months may be a good time to focus on the moats you can build around your business. Your competitors may be distracted by internal issues, and new startups are less likely to get funded or thrive. How can you use this to your advantage?

What does the future bring?
COVID-19 has caused a massive shift in the way people work and live. This creates new opportunities for startups. In general, there are a few ways to segment startups.

a. Growth due to COVID versus Sustainability of growth.
A number of companies have already shown clear outcomes from COVID. For example, Zoom and Instacart will grow rapidly due to COVID and likely reset at a new baseline after the crisis. Other companies will be clearly hurt by the crisis, but this should be temporary. For example, TripActions will undoubtedly snap back once enterprises have people travel again.

The hard part is to tell which companies (a) may or may not sustain their changes in either positive or negative directions (for example, which EdTech co's will snap to a higher baseline and which will not) and (b) which companies that are "doing OK" right now will have their business shrink rapidly once large enterprises have had enough time to react (the next 3-6 months).

b. Capital intensive business versus not.
Capital markets are uncertain right now. Many late stage investors are waiting for the other shoe to drop. For example, big hedge funds with big public market positions (and small private market positions), are spending almost all their time looking for public market mispricing or deals. They are sitting out aspects of late stage private investing right now. The same holds true for some family offices. Traditional late stage VCs are focusing on their own portfolios or cherry picking a handful of startups with the most interest.

The more capital intensive the business, the rougher the next year or so will be. Now is a good time to revisit unit economics, renegotiate with a core supplier, or otherwise determine how to have capital to bridge the next 2-3 years.

Wednesday, March 11, 2020

March 11 Coronavirus (COVID-19) Update For Startups

Last updated: March 11, 2020. Google doc version here.
For Feb 28 doc, please see Google Doc or Blog.

Its been less than two weeks since I posted on COVID-19 / SARS-CoV-2. A lot has happened during that time. Rather than update the old document I decided to write a new one. More background is in the old one. 

Some updates in the last ~2 weeks. This is clearly not “just the flu”.
Given that cases are spreading exponentially, and many areas are undertested, cities or countries can, like Italy, go from a few cases to quarantine in just 2-3 weeks. It should be noted that Wuhan was locked down at 495 cases and 20 deaths. In contrast, Washington State is at 267 cases and 24 deaths, while the Bay Area is quickly growing.

Interventions: Spreading the Epidemiological Curve
In the absence of drugs or a vaccine, non-pharmaceutical interventions like social distancing, the shutting of schools, and quarantines are enacted in epidemics. The most famous studies of social distancing techniques focus on the 1918 flu pandemic.

The key takeaways from these studies include:

Relatedly, the number of hospital beds and ICU capacity matters. Once front line health care workers become tired, exposed, and sick, the system can unravel rapidly. Italy has asked doctors to come out of retirement and graduated nurses early to try to fill the gap.

Much of the goal is to “flatten the curve” - that is spread out the case load of sick people so that the healthcare system does not get overwhelmed to the point of collapse. This prevents both death from COVID-19, but also death from other disease not being treated by a hospital and healthcare staff overwhelmed with an epidemic.

The US has so far not been proactive at epidemiological controls as well as testing and measurement. This leaves it up to individual companies and citizens to act in the general interest of their neighbors until government action is initiated.

Expect More Quarantines
The places with highest risk or both an overwhelmed health care system and quarantine are ones where there is a big up swing in cases and deaths. Given the lack of robust testing in much of the world, case loads are probably dramatically understated. A simple rule of thumb is that if you assume a 1% mortality rate, and 3 weeks for the first people to die, 5 deaths today from COVID means ~500 cases 3 weeks ago. If the case count doubles every week, that means 8*500=4000 cases today. Notably, Washington State has 29 deaths. Assuming we are overcounting cases due to clustering by 2X, we still end up with >10,000 people in Washington state infected, versus the official 290. This math is undoubtedly off, but probably not by 30X.

Given the death and case loads, expect more quarantines to occur. Italy has now quarantined the entire country. Washington State is likely to issue a quarantine. The next most likely candidates for quarantines include France and Spain in Europe and San Francisco Bay Area, NY, and Boston (as a reminder, Wuhan locked down at 495 confirmed cases and 20 deaths)

Table from Austrian site:

Chart above suggests 1-2 weeks before multiple countries hit their own “Italy” moment, at least in terms of cases. Exponents move fast.

If you plan on traveling, you may want to chose a location that you would not mind suddenly finding yourself quarantined in for a few weeks or more.

What about the countries with few cases?
A number of countries claim few to no cases of COVID-19, including Indonesia, India, Russia and Thailand. These countries have taken limited actions to suppress spread or cause social distancing. This suggests a number of these countries may have uncontrolled community spread on their hands which is about to expand dramatically.

Impact To Elderly and Children (and Adults!)

There is an unfortunate meme that COVID-19 is “just the flu”. The reality is, that for 80-85% of people it does indeed seem to be a flu like disease. Unfortunately, around 10% of people end up in the hospital, and 1-3% in the ICU. This is dramatically worse then the flu and causes healthcare resources to get depleted, leading to excess deaths. Above is data from the Chinese CDC (biased undoubtedly due to Wuhan) versus US CDC on flu versus COVID-19. In general, the disease seems to be most aggressive in elderly. However, the case fatality rate in younger adults may still be 2-10X that of the flu.

Below is an interesting graph showing the trade off between event size and likelihood of someone infected with SARS-COV2 being in attendance. Biogen recently held a 175 person event in Boston, which is now tied to 70 cases of COVID-19 including 25+ of the attendees. 

This suggests it is worth canceling most events above 100 people. It also suggests that things like cruise ships, Disney Amusement Parks, marathons, parades, and concerts might not be the best things to pursue right now.

So, What Should My Startup Do? Part 1: Protecting Employees
  • Move to remote working or work from home (WFH).
    • Companies such as Google, Microsoft, Twitter have adopted WFH nationally. This is meant to both protect employees from illness, as well as help the communities in which employees live. Companies are adopting social distancing policies as many governments are determining their course of action.
    • If needed, do first week as a pilot. Try it out, get the kinks out. Fix them, and then try again / keep going. It may take adjustments in processes or tooling to be effective as a work from home startup.
    • Schools will likely shut down in many parts of the US (300 million students are already out in many parts of the world) so many employees will need to be home to take care of kids as soon as that happens. At some point, the local or state government may mandate WFH. Expect this to last 4-12 weeks.
    • Not every company can do WFH without shutting down. For example if you have a biology lab, you will either need to shut all progress or move to a skeleton crew. 
    • Not every employee can WFH. You may need to reserve space for people unable to work from home. If possible (1) space people out at least 6 feet (2 meters). (2) remove communal food and eating. People should either bring in their own food or get individually wrapped food if possible. (3) ensure proper sanitation of work place and availability of cleaners for hand washing. This may shift if the government gets its act together and mandates work from home for non-location-essential work.
    • Let employees know that “working from a coffee house” is not working from home. This defeats the point of social isolation.
    • Plan for the remote work contingency. If you do not immediately adopt work from home, it is worth planning how your company will work if the virus takes off in your country, or in countries where your employees work. Do you adopt a work from home policy or other approaches? What is the threshold for work-from-home? Coinbase has a guide like this
    • In general if you want to help slow the virus moving to WFH now is best (as long as it does not destroy your business, in which case people will permanently be home - no work!)
  • Encourage hand-washing. You can watch a video here for best practices. You may want to add a few purell dispensers around the office as a reminder.
  • Wipe down work areas regularly. There is some evidence suggesting the virus may stick around for at least a few hours, if not more, on surfaces. You may also want to wipe your phone down on a regular basis.
  • Encourage flu vaccination. This will decrease health burden on hospitals and also prevent people from getting the flu and thinking they have COVID-19.
  • Curtail travel and conferences and move to video calls. Your employees may generally want to cease travel and in particular avoid countries where either COVID-19 has started to spread (China, Hong Kong, Iran, Italy, Japan, Korea, Singapore, Thailand).
  • Curtail visitors from other countries. A number of companies are starting to adopt a “no visitor” policy and moved such meetings to a video call.
  • Cancel events. It is wise to cancel group company events. Biogen recently held a 175 person event in Boston, which is now tied to 70 cases of COVID-19 including 25+ of the attendees.
  • Paid sick leave and zero tolerance sick policy. You may ensure all employees have paid sick leave. This is especially important if you still have people in the office and do not want sick employees coming in to make their payday.
  • Hourly workers. Some companies like Microsoft have continued to pay hourly workers displaced by WFH. You may or may not be able to afford this as a company.

So, What Should My Startup Do? Part 2: Dealing With A Downturn
Sequoia Capital has created a lucid guide to the coming economic storm, and how to weather it as a startup. Andressen Horowitz has also added some resources, as has General Catalyst.

Key takeaways:
  • Make your cash last. Imagine the global economy is now in a recession and it will be another 6-12 months before enough people are infected for the virus to go away. Do you have enough money to last until mid-2021 before fundraising if you need to?  It is always possible the economy will escape unscathed. It is best to prepare for this not being the case.
  • Expect slowing growth. Customers may cancel deals last minute or take longer to close. For many startups, sales will slow. If you were growing 3X a year, you may be down to 1.5 or 2X. Can you front load contracts and payments, or find other ways to make up for lost customer in terms of locking in cash? 
  • Can you uniquely win right now? Alternatively is your business uniquely suited to this environment? Can you buy a competitor, accelerate growth, or make the situation workable?
  • Raise money. If you need to, raise money. You may want to do it as a flat round or small bump to close money rapidly. Are there people who wanted in on the last round you can include now? The markets may recover quickly. Alternatively, valuations may be on a slide for the next 6 months. It is often better to have cash than over optimize.

Some resources for startups:

Open Questions
Flu and colds tend to die down in warmer, humid weather. Given the R0 is high for this disease and the human population largely naive (i.e. has never seen this disease), will weather make a difference? Recent modeling paper here (note this is not peer reviewed). Alternatively, will this die down in the northern hemisphere with weather, only to flip to the southern hemisphere for their winter, to return for round 2 in the northern fall and winter? Importantly for the world economy - will China see a second wave?

It is notable that many warm places (UAE, Bahrain, Qatar, Malaysia) are all reporting more cases per capita than the US despite hotter weather. Given the undertesting everywhere, it is hard to tease out reality.

In many epidemics disease course follows two waves. In wave one, an initial infection happens followed by warm weather, governments tightening movements, shutting schools, and in general decreasing the spread of the diseases. Controls are eventually relaxed (people need to work, kids need to go to school etc.) or the weather changes, and then a few months later a second wave of the disease hits and infects a subset of the people who were not infected in the first wave. Eventually, enough people get sick, develop antibodies, and there is a strong enough herd immunity in the population to decrease future out breaks in size.

1918 Spanish Flu had two predominant waves of virus spread.

C. Are any numbers accurate?
An open questions is the accuracy of COVID-19 case numbers. For example, Chinese cases numbers have been thought by some epidemiologists early on to be understate by up to 10X. The US is clearly undertesting. What was the real case load? What is the real denominator to the disease?

D. When will US test sufficiently?
South Korea controlled the outbreak via aggressive testing + social distancing. The US only tested 4500 people as of last week, while South Korea is testing 10,000 a day. While test capacity is starting to ramp, many people are still complaining of a lack of testing for obviously sick patients. At what point will the US truly accelerate testing?

E. How will this impact US politics?
As noted above, COVID-19 is most severe in elderly, although a number of young adults have also been affected. Given the age of the political class, a number of French members of parliament have been hospitalized, the Head of Italy’s Democratic Party and the Health Minister of the UK have all caught the virus. 

At what point does this jump to the US political class? The average age of a US senator is almost 58, with almost 50% of senators over 65. Assuming a number of them catch COVID-19, what are the implications? (I am of course hoping none of them sicken and the country moves aggressively before such a terrible thing happens).

Relatedly, will this impact the election? For example, all the candidates of note are over 70. Will any of them catch COVID-19? Can they still hold events and rallies? Will elderly stay away from polls leading to favoring of a candidate that young voters prefer? This is truly a black swan year. (Or perhaps, more fittingly, a black bat year)