Monday, January 6, 2020

Products I Wish Existed, 2020 Edition

Below is a list of products or startups that I am keeping an eye out for. I skipped listing things I wish existed that can *not* be implemented with today's technology in strong form - for example, the Young Lady's Illustrated Primer - although weak forms of this may be possible today.

A number of these may also turn out to be terrible businesses, and this is not meant as a "trends" or "what's hot" post. Rather, I would adopt a number of the products myself.

A. Software.
1. Trinet for full-time remote/distributed workers.
With the move to more and more remote workers and teams a number of companies are struggling with basic payroll and benefits for non-US employees. This is particularly true if the company is a startup or has many small teams around the world. There is a lot of complexity in terms of whether someone is an independent contractor or not, setting up foreign subsidiaries and payments, working with a PEO etc. The product would be a modern, easy to use way to hire and pay remote full time (and part-time) teams. You can think of it as "Trinet v2". Some of the complexity in the product may be driven by the need to establish foreign entities to hire remote workers. Deel is doing a subset for contractors, while RunaHR is doing payroll in LATAM.

2. New social network.
TikTok is a good reminder for the generational turnover of social products. While most social networks have not proven to be durable over time (Friendster, MySpace, Orkut, ....) the ones that take over generational shifts or new content types and formats have become massive companies. In general, most of the value of a social product is the network and interactions of the people using the product, with the remainder of the value coming from the product software scaffolding itself.

Places to watch for a new social products may include:
  • Generational shifts. Every 5-7 years there is room for a new social network to emerge in the next wave of high school and college students. TikTok is a good example of this, following SnapChat in the past.
  • New content and formats. The rise of podcasting and transcription, short form content highlighting, or other approaches may yield the right engagement model for a new social product to emerge. Past versions of this include Instagram for mobile photos and Pinterest for pins.
  • Gaming communities. DiscordTwitchFortniteMinecraft and others have shown the appetite for social networks in the context of gaming. It is always possible a game or gaming-centric network will make the jump to a broadbased one, assuming its features evolve over time to subsume additional use cases.  
  • Verticals. There is likely room for additional vertical social networks for very large verticals. Neighborhoods in particular still feel like a large potential opportunity over time with the sets of layered information and commerce that can be attached. Nextdoor feels a bit like the Myspace of its vertical in that someone could challenge it - however no Facebook of neighborhoods has arisen to challenge it.
  • High user fatigue. In parallel, user fatigue around spam, bots, and outrage culture without any controls built to deal with this may create the opening for a new network. What would be a network which allowed for more thoughtful discourse? Or at least the ability to more actively mute topics, threads, and groups of users while surfacing better content algorithmically?
  • VR. This is not near term, but a Black Sun like network could emerge eventually with a rise in VR.
  • Past examples relevant for today? Things like IRC, the WELL, and others captured user behavior in unique ways. One aspect was the highly curated or self organized nature of these communities prior to the Internet being deluged by everyone. There is room for a throwback to these more concentrated, in some sense high end intellectual (and non intellectual) communities.
Of the networks that has the cleanest stand alone use case and potential to be worth $100B, the ones which may be displaced or competed with soonest may include (1) Twitter, which captures both high velocity news, public discourse, and CRM in a single communication and media network. Twitter has not evolved much as a product with basics still missing a decade plus later. (2) LinkedIn. Github should have been a strong competitor to LinkedIn for engineers and designers. There may be ways to try to take over other specific use cases from the company as a starting point.

3. More SaaS and Dev Tools.
This one is kind of obvious, and I suggest analysis to be done below to find opportunities in this area (see "Software areas for exploration). I have specific ideas I may try to help pull together or test in 2020.

4. ADT 2.0: Digital neighborhood watch.
The common approach to security cameras is to sell the camera as a DIY installation hardware sales and then charge an attach to cloud storage. However, the traditional path to home security includes services-centric companies like ADT. ADT will install the hardware for you, have a person in the middle in case needed for monitoring, and spins off cash like crazy due to its subscription model.

I would love to see a modern next-generation consumer-centric ADT-like model. This model would include:
  • Services model. Part of the purchase includes a technician showing up and doing the installation. There would be no DIY component.
  • Subscription only. The product would be a hardware/software/services bundle with an up front fee and ongoing subscription as default.
  • Network driven. All cameras in a neighborhood would be virtually networked - for example, as a user I could report that someone stole packages from my doorstep and an alert would go out to all my neighbors to their phones to bring their own packages inside. There are interesting network effects available here as well as the potential for a neighborhood social product to be built on top.
  • Data exhaust. If you have enough camera density by neighborhood, other things suddenly become interesting adjacent businesses. For example, could you add a pollution sensor to installs and get unique and valuable data? Weather data? There are lots of opportunities here outside of home security once you have an installed base.
I have long wanted to fund or incubate a company in this area. Nest and Ring are trying to do things on the consumer side, but not quite there yet. A company to watch in an adjacent area is Flock Safety.

5. Neighborhood (and corporate?) pollution sensor networks.
Relatedly, the California wildfires have caused me as a parent to look frequently at pollution data. Even basal pollution levels may impact cognitive function. There seems to be room for both home/neighborhood and corporate level pollution sensor networks. One could imagine a service where companies or people buy hardware + analytics for their homes or facilities and then a subscription for people who do not own the hardware but live nearby. An early player in this area is Aclima which has done cool things with Google.

6. Climate product bundle for enterprises.
A bundled climate change product could be sold to enterprises including carbon offsets, carbon sequestration, solar installations, and other climate services. A number of technology companies adopted climate products early, and given employee interest in the area could lead to faster sales cycle then the past.

As an aside, a friend of mine and I put up a consumer carbon offset site back in 2007 due to concerns with global warming. The idea was to help people buy carbon offsets for their own activities. At the time, we saw very little interest.

As an additional aside, Nuclear is a strong potential solution for climate change but is politically unpopular. Could something be done to address this[1]?

7. Software-only defense contractor.
Anduril is a next-generation defense contractor focused on hardware + software (including machine learning) bundles. There is probably room for an adjacent company focused solely on software-centric machine learning and intelligence applications. Unlike Palantir, which seems to have a services-centric model, this new company would be focused on a more SaaS-centric software-driven model. The rise of machine learning and machine vision as analytical tools opens up the door for a new contractor to emerge to challenge incumbents, and also to provide new applications not available in the pre-ML world.

B. Software areas for exploration:
8. Global 2000 company IT needs.
Silicon Valley enterprise startups often focus on the needs of other mid-stage tech companies as customers, rather than the large enterprises that make up much of the business world by revenue. There are probably a dozen $10B+ enterprise software companies to build that are unique to the largest companies of the world that most Silicon Valley founders are never exposed to (UIPath is a good example of a company serving such a need).

I would love to see the following analysis:
  • A map of repetitious tasks, spreadsheets, and manual data extraction by function in the Fortune 500.
  • Budget breakdown of current software spend, by function, by line, in the Fortune 500.
  • A view of what Accenture, CapGemini, and Deloitte keep building over and over for large enterprises. Undoubtedly a subset of these custom consulting projects can be turned into SaaS software.
  • A tougher analysis to do is to ask what internal software projects various tech companies are working on. If you can get the list from 3-4 companies, you will undoubtedly see a few internal tool or product examples that should be built as a SaaS product for everyone.
9. More RPA and NoCode.
RPA (basically tools anyone can use to create programmatic APIs on top of legacy software) and NoCode/LowCode (letting anyone build an app with a spreadsheet as a database representation) are the biggest shifts in enterprise software since the cloud. In a few years, aspects of RPA and NoCode are likely to be incorporated into an increasing proportion of enterprise applications as a standard tool set. While the early leading apps of this area (e.g. UIPathAirtable, ...) make great headways, additional players like NotionParabolaRetool, and Simpo continue to take on additional uses cases. There is a lot more room for new players tackling specific use cases or verticals, or add on apps like Rossum or Instabase for NLP and data extraction into NoCode tables and apps like Airtable, or RPA apps like UIPath.

NoCode and RPA are likely to become more verticalized in terms of both function that uses them, and market served. More companies in these markets will emerge in 2020 and beyond.

C. Biotech.
Although I am no longer helping to start or invest in biotech (outside of my existing involvements), the following strike me as important.

10. Fertility for everyone (and perhaps saving endangered species in the future)
It is now possible to take a skin cell from a mouse, turn it into an induced stem cell, and then convert that stem cell into a sperm or egg. The implications to society if someone were to translate this to humans is dramatic. Imagine if any person, of any age, could have a child. Women could decide to push back child bearing dramatically without freezing eggs in advance. Gay couples could have kids together with genes from both sides. There are also nefarious uses - for example if you surreptitiously scrape skin off Kanye's hands in a club you might be able to show up 9 months later with a baby and paternity suit. Ovid Research is working on one aspect of this.

Relatedly, for endangered species, a non-commercial effort should be made to freeze cells from every living animal (assuming non-disruptive) and also to create embryonic stem cell or germline lines for future use in potential species rescue or rebirth. Here is a white rhino example of potential proof of concept approaches.

11. Longevity basics
Right now the bar for longevity or anti-aging companies is a high one (cure aging!). While each company starts with a specific targeted indication, I sometimes wonder if people are being too bold in application versus ambition. It would be great to want to cure aging, but start really small. The bar for anti-aging may be too high and too science-centric. Are there simple approaches to help people rapidly with off the shelf or repositioned drugs?

I think companies could set a lower bar and start by restoring function that dramatically improves lots of people's quality of life. Examples would include eyesight and hearing. For example, many people start to get blurry vision in their 40s, due to a variety of factors including the mix of proteins in the lens of the eye causing hardening and potentially the muscle holding the lens aging. Could a simple approach like rapamycin droplets in the eye reverse aspects of aging and therefore eyesight? Similarly, are there treatments that could restore lost hair in men or have other cosmetic applications?

As an aside, if you want to join an anti-aging company, BioAge and Spring Discovery are the two to watch IMO (I am involved with both, as well as a number of other companies mentioned in this post).

D. Biotech Areas For Exploration:
12. Biomarkers for everything (wetware and/or AI and digital)
Being able to measure a physiological change or response allows you to both understand a disease and its progression, as well as to create drugs to treat the disease. Major conditions such as Alzheimer's disease have traditionally been limited in drug creation due to an inability to find biomarkers that correspond with disease progression. This means that historically Alzheimer's trials have only been able to recruit people with late stages of the disease for trials. It would be useful to do deep proteomic, metabolomic, and potentially methalomic analysis of people with different diseases to start to develop robust biomarkers for all major diseases.

There are also "electronic" biomarkers that should be developed and put into use. For example, by using machine learning on pictures or short videos of children can you tell where they might be on the autism spectrum? Or can you tell if someone is developing Alzheimer's or other disorders based on changes in gait and phone sensors? There is a wealth of digital data we are not using for diagnosis and treatment. The challenge is taking a one-off research study and translating it into an actual product in widespread use.

Exact Sciences is a good example of a colon cancer biomarker that has decreased the needs for colonoscopies while creating a $10 billion+ company. Unfortunately, like everything in healthcare, the road from a great biomarker to commercial success may be long and difficult. Exact Sciences was founded in 1995, when public in 2001, and had a flat market cap for about a decade before in-licensing and developing its colon cancer screen in 2009-2014. A better product alone in healthcare does not mean success or adoption without the right industry body approvals and insurance reimbursement. However, from a human impact perspective this field holds enormous potential.

NOTES
[1] For nuclear - an update on the decades old technology used in current reactors (for example molten core reactors or SMRs) + a PR campaign targeting millennials and GenZ seems valuable to do. Maybe you rebrand the newer safer technology something completely unrelated to nuclear? "Green atomics" or the like? When I last looked at nuclear 5-6 years ago, a big cost to build and deploy was largely NIMBYism driven - cost delays due to protests led to uncertain completion timelines, delays, and financing costs. Am upgrade to current technology + PR campaign might help here.



    RELATED POSTS:Markets:

    Friday, January 3, 2020

    5 SaaS Features For Big Enterprises

    In order to sell to enterprises, SaaS startups usually need to add the same 5 core product features. These features allow for a dramatic increase in price (often a few fold) and the ability to close large accounts that may scale in the millions of dollars per customer.

    The below assumes you have something that is working bottoms up that you want to adapt for enterprise. If your core product does not have product/market fit, adding the below features won't help much.

    Enterprise features usually include:

    1. Single Sign On (SSO).
    In some cases, SSO is sufficient to cause enterprises to adopt your product. This is often the minimal enterprise feature set.

    2. Admin View & Permissions.
    Often you want to be able to have an admin set permissions for individuals or groups of users. Alternatively, you want to be able to set preferences at the group level versus individual level (for example, you do not want to add every new engineering hire to 50 slack channels. You would rather add them to "eng" and have it propagate to all channels. Lastly, you want to be able to have an admin take over or manage specific permissions. This is often the second or thing most requested feature for enterprises.

    3. Security & Compliance.
    The most common compliance request is SOC2, which sets the data and information security policies, procedures and approaches. Other forms of compliance may include GDPR, HIPAA (for healthcare companies) and FEDRamp (for US Government contractors). In general, these security compliance guidelines set policies around things like data encryption, rules for which employees have access to what data, and even things like physical security of servers on which data is stored.

    Enterprises may additionally request a separate security audit or review, or may want their security team to come visit and ask questions first hand.

    4. Invoicing.
    Your bottom up pricing often allows an early adopter to put your product on a credit card. For large enterprises there are specific types of invoices / integrations you may need to do over time. This is usually not a deal breaker, but makes it easier to get paid.

    5. Paperwork.
    Additionally, you may need to do a lot of paperwork and procurement reviews with large enterprises. This may necessitate a sales operations or support team. Not a product feature, but good to know about.
    Honorable Mention - On Premise.
    As an aside - an honorable mention for enterprise sales may be on premise software - a subset of the world does not want its code, communications, or other sensitive data hosted on a third party providers cloud, or with a startups security know how. The world largely continues to move to the cloud, but sometimes this is necessary for at least a subset of customers. Some SaaS companies chose to skip customers who require an on-premise solution, while for others the business case / revenue is sufficient to justify it.

    Generally the 5 features listed above are straight forward to do and take at most a few engineers for a few months each (and may be banged out by a small team over 3-9 months depending on how many are needed, and the complexity of e.g. admin view). Once you have these 5 features core procurement and common use cases should not be a major obstacle to adoption (although product-specific features may still always be an issue. 

    As an aside, the ability to increase pricing dramatically with even a subset of these features often surprises founders. :)

    MY BOOK
    RELATED POSTS
    Sales & BD:
    Hiring:

    Markets:

    Managing Investors


    Wednesday, December 18, 2019

    Magic Startup Moments

    Startups are hard. There are a few moments in every startups life that feel magical and make the grind (at least temporarily) worth it. Magic startup moments include:

    Your first customer payment.
    It is not a coincidence that so many restaurants have their first $10 bill framed and hung on a wall. Finding someone who wants your product enough to pay for it, and making money from something you built with your own two hands feels good.

    Your first magic executive.
    As a startup matures, its leadership bench gets built out. Many first-time founders of managers tend to fight delegating to others. Or they promote "smart, junior athletes" to run a function like engineering or sales and watch them flounder (often while thinking the floundering is normal execution). Then, that magic moment happens - you hire a tried and true, excellent executive for a part of the company. Suddenly, great people get hired, the right initiatives are launched, and that area of the company starts to function incredibly well. This is often the moment when founders learn to delegate, and think "I want more people like her".

    That special customer thank you.
    Many companies change the lives of their customers for the better without realizing it. For mission-centric companies (for example a genomics company) deep, from-the-heart thank you notes from customers may arrive quite early. For other types of companies a thoughtful email or letter will suddenly show up thanking you for helping your customer and for making their lives better. These letters give you an incentive to keep going.

    The sale.
    The ends of things are often bittersweet. Successfully selling your company and feeling that you finally made it, that your employees have a stable place to work, that you made money for yourself, your family, and your investors, and that you can relax just a little bit, can be a truly magic moment.

    The reunion.
    Sometimes an employee will reach out to you years after they left your company to thank you for the formative experience they had working for you. Or, a bunch of "old timers" from the company's early days get together to swap stories and discuss what was and the crazy things your company did to ship that product, keep the lights on, or land that partnership.



    RELATED POSTSM&A:


    Startup life:

    Markets:
    Raising Money

    Managing Investors

    Tuesday, October 1, 2019

    Fear of Sales

    "Our direct sales and customer success efforts are focused on larger organizations who have a greater number of users and teams and have the potential to increase spend over time... We had 575 Paid Customers >$100,000 of ARR as of January 31, 2019, which accounted for approximately 40% of our revenue in fiscal year 2019." 
    -From the Slack S1

    We have a unique model that combines viral enthusiasm for our platform with a multipronged go-to-market strategy for optimal efficiency... Our sales model allows us to efficiently turn a single non-paying user into a full enterprise deployment. For the fiscal year ended January 31, 2019, 55% of our 344 customers that contributed more than $100,000 of revenue started with at least one free host prior to subscribing. These 344 customers contributed 30% of revenue in the fiscal year ended January 31, 2019. 
    -From the Zoom S1

    Two of the most viral, bottoms-up products in the history of enterprise software benefit immensely from direct sales[0]. For Slack, 40% of their revenue comes from just 575 customers (out of 88,000 paying customers and 500,000 total customers!) while for Zoom 344 customers generate 30% of their revenue (up from 22% and 25% in the prior two years). Big customers are growing faster for Zoom versus customers overall. I know Slack started on enterprise sales late, so it is possible this number could have been even larger today.

    Many technical or product-centric founders have an innate (and inappropriate) fear of sales[0]. While it is inevitable that their company will need to adopt direct sales, technical founders often procrastinate on building a sales force. This is usually a mistake. If direct sales could boost growth from 15% a month to 20% a quarter on $1M in ARR, the company will have $184M in revenue versus $93M in revenue after 4 years[1].


    If the company actually does not have good bottoms-up adoption, the impact sales can be even more striking. In DataDog's case, 72% of their ARR is from large enterprises.
    We have a highly efficient go-to-market model, which consists of a self-service tier, a high velocity inside sales team, and an enterprise sales force. As of June 30, 2019, we had approximately 8,800 customers... Approximately 590 of our customers as of June 30, 2019 had annual run-rate revenue, or ARR, of $100,000 or more... accounting for approximately 72%of our ARR... Further, as of June 30, 2019, we had 42 customers with ARR of $1.0 million or more" 
    - DataDog S1 

    So why do product-centric founders fear sales?
    1. Familiarity.
    Most technical founders have strong familiarity with bottoms-up sales. They download open source packages all the time, and adopted Github or PagerDuty or Stripe through word of mouth. What many founders have not seen directly is how well a repeatable sales engine can work.

    The simplest advice to founders on this point is "just do it". Just as you needed to learn how to do PR,  how to hire and negotiate salaries, how to raise money, or how to hire other specific functions (finance, legal, customer support etc, you will need to learn how to do hire sales people and add go-to-market as a function.

    To learn how to hire a sales leader, you need to take the same steps you took to learn how to hire marketing or customer support or legal - ask others who are 2-3 years ahead you as founders who have done something similar how they did it. Or go and find 3-5 great sales leaders or CROs and ask them how to hire, and interview for the role, and how to best run the function.

    Just as you do not need to run your finance team yourself, you can hire a great sales leader to partner with to drive sales for you.

    2. Perception of sales as an art versus science. In reality there is a sales playbook.
    Since sales are stereotypically driven by people with liberal arts degrees who played sportsball in college, technical founders often view it as a bit of dark magic involving expensive dinners, cross country flights, and a lot of small talk, high-fiving, and gong-hitting.

    In reality, sales is a giant process engineering exercise which has benefited from 40 years of iteration. A great sales leader will have an effective playbook they can apply, a pipeline process to run, and a hiring and compensation approach that will work for your market.

    Sales at its best is a repeatable, understandable operational process that can be run in a metric driven way. Indeed, sales productivity is dramatically easier to track and understand versus engineering productivity. Metrics may include revenue productivity per head, payback periods, and funnel or pilot conversion and close.

    The temptation of a technical founder is to make sales up as they go along. For your sake, it is best to resist this temptation to re-invent the wheel and to hire someone who knows what they are doing. You will obviously need to work with this individual to set overall objectives, interview the first hires for culture and the like. But you should view sales as an integral part of your startup team - and not something to design yourself from scratch.

    3. Cool factor.
    Bottoms-up adoption is considered much cooler and more organic than doing inside or direct sales. Your customers are coming to you. How cool is that? Why would you ever lower yourself by going to them and actively engaging them?

    The even cooler thing is to build a monster company having enormous impact and a platform you can use to influence the lives of millions and generate the revenue and market cap to affect humanity-level productivity and social change.

    If resistance comes from your team versus you, explain to them that your goal as a company is to build something great. To build something at scale, you will need to get big companies to adopt your product (even more people can use it!), that all the companies your employees use and admire actually have sales forces, and that sales at its best can provide insights that would be more challenging to get otherwise (see below)

    4. Cultural concerns.
    "Sales will be an invasion of fratty liberal arts people and our culture will be ruined! There will be people with nicely pressed shirts making small talk in loud voices while making too much direct eye contact! Don't they realize life is not a staring contest?!?!!!"[2]

    Often when I hear about the cultural concerns of working with sales, it is from people who have never worked with sales. The reality is:
    a. You will ramp sales headcount slowly at first. Unless you have built a sales org before, it is wise to start small. Add a sales leader. See how it goes. Have her add 1-2 people. See how it goes.
    b. Sales tends to be geographically distributed. Direct sales people often work in region since much of their time is on the road. Inside sales teams (who are often phone and email based sales) may be more cost efficient to have in a remote office. The reality is you will eventually be going out of your way to get the sales people together in one spot for offsites and conferences, or to spend time at HQ, versus them taking over your home office. The primary place I have seen sales actually dominate a culture is when you have a remote sales office and then add a handful of engineers and leave the engineers to fend for themselves. Or when you have a sales-centric CEO to begin with, in which case the culture will be sales-centric as expected.

    Sales will change your culture, but it is an inevitable change. Instead of fighting it determine how you can productively incorporate sales people. How can you grow in a sustainable manner? How do you want your culture to evolve? Taking it slow is a better long term solution than not doing it at all (or waiting to long and regretting it).

    5. Sales at its best.
    At its best, sales is not only a large revenue driver, but also a way to get customer feedback and input back to the team. It should drive rapid revenue growth while feeding your team competitive intelligence, customer concerns, feature requests, and other insights from the field. These can help drive business strategy, product roadmaps, and your ultimate success as a company.

    6. Sales can cause a company to overspend and crash & burn.
    I have seen instances where a company overhires a sales force before it has a repeatable working sales model, or real product-market fit. In this case, a sales team can be existentially bad for a company. You will burn through all your cash before sales shows success. Often this happens when an incoming sales leader over promises and over hires, and sales productivity is slower than expected. For example, a company may hire 20 sales reps and promise a 6 months payback period, when in reality the new sales reps ramp slower and are less productive. Suddenly, the company is stuck paying out a sales guarantee to new hires and it blows up due to poor cash management. This pattern is actually quite common. It is often better to start sales teams small and experimental, and then ramp.

    Give yourself permission to ramp sales slowly, make sure things are working, and then ramp some more.

    Fundamental Decision
    The fundamental decision a technical or product-centric founder faces is this:
    Do you want your product to be adopted by as many people as possible, including people working at the largest companies in the world? Do you want to win your market and industry?

    If your answer is yes and you are running a SaaS or enterprise company, you should start to build a sales force as soon as you have signs of product market fit[3]. Don't wait. Don't procrastinate. Don't make excuses. Even if bottoms-up is working, you will end up with a smaller, less important company if you wait on building real sales and customer success teams.

    You should feel free to start small and take incremental steps - but the key is to take the steps early. Don't wait too long. You may also go through 1-2 sales leaders before finding the right one (and understanding what to look for for your product and culture)[4]. That is OK too. Give yourself permission to make a mistake or two, and let your team know this is all an experiment. This is all about the long term.

    Thanks to Stephanie (Schatz) Friedman for early comments on this post.

    NOTES
    [0] A quick aside on sales definitions:


    [1] You of course want good unit economics and payback times on any customer acquisition efforts, including direct sales.

    [2] I am of course joking. I have a technical Ph.D. and a math degree so actually kind of hate most small talk.

    [3] Assuming your price point and return on a customer is high enough. For example, if you are only making $100 per year per customer, you can not afford a sales force. At tens of thousands per customer, you can afford an inside sales force. At $100K+ per customer, you can afford direct field sales.

    [4] By the way, if your gut is telling you the new sales leader or AE is doing a bad job, they probably are. I have often seen technical or product-centric founders second guess their own judgement on people in these roles. Excellence has the same feel irrespective of function.