The New Virtual Accelerator

The early stage venture environment has changed enormously in the last 10 years. It has never been easier to start a venture and access seed capital; nor, arguably, has it been harder to separate ahead of the pack in the overcrowded startup world, which may have reached Peak MicroVC and Peak Accelerator. Seed only funds barely existed ten years ago, but now number around 300. Likewise with accelerators, which have gone from single digits to over 1000 in the last decade. The numbers have exploded, creating a staggeringly noisy environment for entrepreneurs, investors and customers alike. How can any program create real value? As we learned in The Invincibles, “When everyone’s super, no one will be.”

While I am not predicting any time frame for bubbles (although it IS looking rather, shall we say, “carbonated” out there), it is inevitable that there will be major consolidation in accelerators and early stage investors, and that that will be healthy for Darwinian reasons in the long-term. This is the first in a series of articles that will strip down the “venture stack” to its three main components–knowledge, connections and money–to discuss how the startup world will change. In this first article we will look at accelerators, and predict how the top accelerators will evolve to offer a superior product while scaling successfully via the use of software. It ends with a hypothetical example of a new accelerator based on real people and institutions. Follow-on posts will look at aggregating startup business knowledge in a public, Wikipedia-like framework as well as a third post focusing on managing investor and mentor networks. We’ll conclude by revisiting seed funding economics and mechanisms, including more specific predictions how hybrid investor/accelerators (such as MIT’s The Engine) will flourish.

Seed Accelerators Today

Marc Andreessen famously predicted that software will eat the world. Yet with very few exceptions, like AngelList, the VC world itself has changed relatively little in the past 30 years. The same holds true for accelerators–we see few accelerators straying from the original Y Combinator model dating back 11 years: almost all are locked into the same 3–4 month program offering teams modest monetary investments for 5–10% stakes, covering the same basic Startup 101 principles, preaching the same gospel, and limiting themselves to a physical location processing a batch of companies simultaneously. Virtually all culminate in some form of “Demo Day” (which strangely rarely show an actual demo) to raise funds. Talk about irony–where is the disruption and innovation in the last 10 years while the number of accelerators have increased more than two orders of magnitude?!

Nevertheless, the best accelerators already have started making significant changes, and a new and better model is emerging. The next generation of accelerators will be predominantly or totally virtual; top programs already embracing online programs. Techstars is running its first virtual accelerator class called “Techstars Anywhere”. In a few weeks, YCombinator will open-source some of their lectures and content via a “Startup School” MOOC. Already virtual Demo Days, notably the one from StartX, are making inroads on fundraising, and online learning programs have been established. To take off, only networking needs to be better integrated online in order to rebundle the current components to create the new virtual accelerator.

With rare exceptions, the top accelerators are for-profit– as a business model, spreading equity over a wide sampling of promising, highly optional companies is hard to beat. And like all successful business models, those accelerators want to scale profitability. In order to increase the number of companies participating in accelerators while keeping product quality high, going virtual is inevitable. While there will always be a place for small size, high quality, artisanal style programs (e.g., AngelPad), the best-known players are ramping up in size, graduating 150–250 companies a year. Scale like this requires a veritable “plug and play” modular system, with continuous or “on-demand” delivery of knowledge, access to networks and money. Such a system is massively data-driven and software-enabled. And when we get there, Andreessen’s law finally will have arrived in startup land: only then can software add accelerators to its list of eaten institutions.

The Venture Stack: Capital, Networks, and Data

In The Tipping Point, Malcolm Gladwell talks about the necessary elements to go beyond ideas and modest traction to overwhelming success. A successful team requires a few superstars who can be characterized as salesmen, connectors, and mavens. In the same way, we can look at all of venture (whether it be a startup, an accelerator, or an investment entity) as being increasingly valuable the more they possess capability in the three areas that correspond to Gladwell’s triumvirate: money, networks, and knowledge. We’ve all heard about “Dumb Money”–when money brings nothing else to the party. So, how can accelerators assemble all three? Furthermore, how can they do so on a custom basis to suit the needs of their startups?

Currently, the default mode is like the traditional grade school method–all companies cover the same material simultaneously, experience the same exposure to mentors and investors at the same time, and fundraise at the same time. While this is easiest for those running the program, it is not necessarily (or even probably) the best path for the companies. But what if we could apply all the magic on a when-needed basis, and let those who can skip ahead do so, while letting those behind take their time without fears of missing the all-important Demo Day? While the necessary ingredients–money, network and knowlege–remain constant, the recipe should vary by company. Is, say, a bounded 3 month program appropriate for a medical device company that may require years to go through trials and FDA approvals? But if the process varies, how can an accelerator administer differing solutions? The answer lies in data, software, and program design.

Solution: The Functional Approach

In Understanding Abundance: The Next Big Thing, Alex Danco argues that the Next Big Thing will be companies that sell functions, not objects. Want transportation? Don’t buy an underutilized object (a car), but a more cost-effective function (getting somewhere via Uber). Want computing power? Don’t build your own server farm, rent out cycles from Amazon. Software enables the ability to access functions without investing in objects. With this in mind, let’s not try to assemble a bricks and mortar accelerator, but simply deliver the functions startups require. What kind of functions should the new accelerator be able to deliver? Almost anything an entrepreneur needs, throughout the life cycle: fundraising, recruiting, basic legal and admin, coding, manufacturing, distribution, design, PR, marketing…you get the idea.

Accessing Knowledge–What to Learn versus What to Rent

The stereotypical hot young startup consists of a few coders with hoodies–or perhaps a scientist sitting in a wet lab with an earthshaking invention. The startup at this stage isn’t much more than some specialized knowledge, intuition, a killer prototype, and a lot of drive. Why do most fail? It’s not for lack of money nowadays, it’s lack of basic knowledge. Sometimes this need is even more pedestrian than how to access money or connections, or how to deal with co-founders. (By the way, I find most accelerators negligent on providing hands-on executive coaching regarding personnel issues.) Almost every young company lacks some of the business basics which have been the traditional curriculum in business schools. Should you expect the young Mark Zuckerberg to have mastered accounting, HR best practices, or legal issues? Of course not. There are two ways to get access to knowledge–learn it, or rent it. Our new model provides for both.

Services as a Differentiated Recruiting Tool

Just as VCs compete by stressing their “value add” beyond writing a check, the best accelerators increasingly will compete in providing access to more basic functions to their companies. Some of the best venture capitalist firms already provide services (shoutout to GV, Social Capital, Andreessen Horowitz) to their teams, whether that be lending in-house teams of designers, growth engineers, or recruiters. Accelerators should do no less, additionally focusing on the back office services overlooked by virtually all investors (except for ff Venture Capital–nice work, folks.) Pedestrian issues like accounting which TOTALLY should be offloaded by startups.

Thus, the rich programs will get richer–if you had to choose between two programs, one being traditional, and the other offering full concierge essential services, where would you go? Some of this happens on a small scale now: many accelerators offer standard legal docs for, say, a financing may be shared, or teams of dedicated associates hired to assist companies, but nothing on a grand scale. To me it’s simple: the percentage of equity, i.e., the cost, of attending a generic accelerator will decrease with competition, but accelerators can not only maintain but increase their margins while attracting better applicants by providing more value through excellent, standardized services. Whether those services should be charged a la carte (at cost) or as prix fixe, I can’t tell, but an upgrading of services and value offered is inevitable. Eventually three levels of services will get offered: 1)DIY apps; 2)a curated suite of apps provided on arrival, all working together and tuned for different types of businesses; and 3) vetted, trusted outside vendors recommended by the accelerator.

Example 1: The Turnkey Software Stack (Levels 1 and 2)

What would attract me to one startup over another? For instance, licenses (free, subsidized or otherwise discounted) for homegrown and external software applications, e.g., CRMs (e.g., Salesforce, NetSuite) project management software (Trello, Asana, Basecamp), HR tools (Namely, SAP), capital table administration (eShares, Shoobx), and bank accounts (Silicon Valley Bank), set up and provided as part of the onboarding process. Already Amazon, Microsoft, Google and the like offer massive amounts of free cloud storage to startups going through accelerators. When setting those up, why not also set up a feed where, say, Google Analytics and AWS data are shared with the accelerator via Domo? Just automate it from Day One into a customizable dashboard. What if the company wedded to their own software? Fine, just provide an API or mirror the data onto the mothership’s platform.

Example 2: Vetted, Discounted Outsourcing (Level 3)

For many teams, access to programmers is a massive challenge. A solution would be to provide credible outsourcing connections. If an accelerator can confidently source, say, external programming (say, from approved dev shops in Ukraine, college interns, or in-house staff), that will make for faster growing companies, hopefully with better resources than they have on their own at this early stage. Of course this isn’t free–but I’ll leave the business model discussion for a followup post.

In summary, the new accelerator provides not only a broad range of basic services, but also acts as a concierge to set up on-demand services at pre-negotiated discounts. If you can spin up (or down) computing services from Amazon, or office space from WeWork, why not startup services sourced via the accelerator?

The Dashboard as a Teaching/Management Tool

In our new ideal accelerator model, we realize that different teams have different needs–but if it’s a large or virtual accelerator, how can we keep track? The answer is linking all of the companies’ data to a dashboard. All companies are encouraged, perhaps required, to submit all relevant data–financial, operational, product- and customer-based information–into a centralized database which informs the dashboard. In doing so, the accelerator, in its role as both coach and service provider, can keep current of the most important needs. Is there less than x months worth of cash? Adjust the game plan. Customer funnels stuck at a certain stage? Investigate and fix. The solutions come from offering standardized software and APIs, creating “Need to Know” permissioning to safeguard companies’ data, and creating rules and alerts that suggest certain solutions and actions, dependent on the data. Some accelerators have invested heavily in software, but no one has offered this type of utility to companies. What dashboards should be built? It depends on the types of companies participating, but here’s one viable view. SaaS companies will clearly have different data needs than hardware companies. The broader the mandate of the accelerator, the more comprehensive the managerial accounting needs to be, but simple financial measures will be common to all. (Free idea: with the massive supply of accelerators and 100x more relevant companies needing this data, there is a great business here for someone to build.)

Combining Dashboards and Services to Determine What Knowledge to Deliver, When

By knowing exactly where a company is in its lifecycle, the proper curriculum or resources can be shared. Learning how to, say, extend a marketplace from a home base to multiple cities is irrelevant for a company with no cash, few employees and other bad metrics. So why would you suggest, say, on-line videos or mentoring on scaling? A dashboard, matched with access to an online curated library and connections to mentors or staff, solves this “what knowledge when” problem. Just as Netflix recommends movies, each stage completed or need uncovered should trigger a suggested menu or curriculum of services available.

Raising Money

Let’s get straight to the heart of the matter for most startups joining accelerators–access to investors and capital. The needs for capital (as well as knowledge) change according to stage. (Note the recent trend of accelerators to offer multi-stage support, e.g., 500 Startups’ distinction between its “Seed Program” and its “Series A Program”, and the follow-on funds assembled to maintain pro-rata holdings with its winners.) But I prefer to breakdown the funding as “traditional”, i.e., a Demo Day which acts as a forcing function, vs. a continuous model–accessing capital as needed. My earlier post “Why Demo Days Have to Change” discusses exactly how software and video can combine to allow accelerators, virtual or traditional, reach the perfect audience. But the more interesting twist, which doesn’t lend itself to rigid 3–4 month programs, is how companies should raise capital before or after a Demo Day, as well as in A, B and later rounds. An accelerator wants its portfolio companies to find the best possible sources of capital; only when they link detailed knowledge of a company’s metrics to a properly mapped investor network will they be able to do so consistently.

Peter Phan of Science persuasively discusses how a large percentage of his job consists of knowing which investors at which firms are the right people to approach for any of his portfolio companies, and that includes when to make the connection. Now, add software capability, filled in by those VCs themselves, which shows what exact metrics and qualities are wanted. “I’m looking for SaaS companies with x% monthly growth in revenues when they reach $Z in MRR”.

There is no reason why an accelerator cannot make these introductions happen–not only limited to the duration the program, but forever after. If the dashboard adds value, alumni will keep feeding you the data. (N.B. Give companies a toggle switch to turn off certain data, so that if they feel the need for privacy, they don’t need to turn off the entire model.) Once the accelerator is knows the needs of both sides of the market, successful transactions happen. More detail on this to come in future posts.

Putting It All Together–The Hypothetical SaaStr Accelerator

Let’s spin up an example as a thought experiment to show how a viable accelerator can be spun up. While I don’t know him, I’m an admirer of what Jason Lemkin has created. Five years ago, this ex-lawyer left Adobe, which had purchased his startup Echosign. After beginning with a modest following built on Quora and then his blog, he created the SaaStr brand. Today he has a massively successful SaaS-specific conference and sponsors a podcast managed by Harry Stebbings; add to the mix an enviable track record as an angel investor and advisor. (This profile, btw, is fairly similar to other celebrity angels like Jason Calacanis–who has launched an accelerator–and Tim Ferriss. Lemkin’s SaaS pedigree among VCs is reminiscent of that of David Skok and Tom Tunguz, neither of which yet has an accelerator–so this example can be played out with many others.)

Look at what the hypothetical SaaStr accelerator brings to the party: 1)name recognition/brand; 2) domain expertise of the founder; 3)access to star mentors/investors/speakers; 4) deep community following (33,000 newsletter subscribers) and 5) money–$70mm raised for his new seed fund, almost all from later stage institutional investors.

Making this a slam dunk, Lemkin also has gotten into the co-working space business, renting out 15,000 sq. feet to SaaS-related companies in SF. To complete this thought experiment, let’s look at how he could create a killer accelerator to complement and augment his role as a seed stage investor, strengthening the ecosystem as well as his investment portfolio.

  1. Create a keiretsu or syndicate-like grouping of his favorite vendors. Pick one or more law firms, accountants/bookkeepers, cloud storage provider, cloud compute provider, recruiting company, project software providers, dev shops, freelance designers, etc. Get them to offer deeply discounted products or free services for 1 year to all companies. After that, services default to market rates.
  2. Hire a few programmers with skills in APIs and middleware to connect up the pieces and feed into the dashboard. (Jason is already a maven and knows every SaaS metric he wants to track.)
  3. Raise funds (either from your current fund, or via LPs) to manage the accelerator. Figure on $75k per company to run it well, plus any cash given to each company in exchange for the x% stake they give up. Create a commingled fund where the participants can offer 1% of their equity in return for a pro-rata share of everyone else’s equity. Such arrangements will benefit 90% of teams (only the most successful don’t benefit economically, and they are happy anyway) to help establish an esprit de corps and collaborative atmosphere.
  4. Guarantee follow-on investments to new companies based on milestones reached. (“If you can do this, we promise to fund you on a published scale valuation, adjusted quarterly to be at market, or follow on other qualified funding.”) Publish those milestones and funding level guidelines.
  5. Send out the word for investors and mentors in SaaS matters, and have them register profiles in a database, most of which will be visible to the teams.
  6. Take a month to design a stage-relevant curriculum. Repurpose existing SaaStr content made over the last 5 years. Add in a few interactive workshop dates that can be taught by mentors or the program managers hired.
  7. Tap the SaaStr community for applicants, hit TechCrunch, etc., and open applications.
  8. Sort the applicants into two piles: local and remote. Take some of both, reserving room available in the new SaaStr Co-Selling Space.
  9. Set up rock-solid video conferencing capabilities. Add off-the-record private AMAs and talks with the mentor and investor networks as added spice. Those that can be publicized will create new content that can be published and distributed to the SaaStr community and beyond.
  10. Start ‘er up! Involve the broad SaaStr community and platform to help companies dogfood their products–while many startups seem dependent on Product Hunt, the SaaStr community has an entire community of beta-testers pre-established even before an alumni network is established.
  11. Reserve some percentage of equity to be earmarked for top mentors, staff, and some business partners (outside service providers, sponsors). Dole it out in consultation with the companies to those who most deserve it.
  12. Like anything else, build/measure/learn. Sweat the details on how to maximize peer involvement and community building, especially for those not in the office. And watch the rocketship launch. Open-source as much content as possible–doing so will result in better applicants and good karma.

Voila–a blueprint for a differentiated, competitive accelerator. For Jason, the end product is a “well-stocked pond” giving him matchless visibility into the viability and potential of the participating startups. He will, of course, need to set rules regarding anti-selection and ownership percentage so as not to drive away other A round investors. But all in all, it offers fantastic economic optionality for Jason, and possibly a terrific experience for companies desiring going through a SaaS vertical accelerator.

Sorry to be so long-winded…but there’s more. Stay tuned for the next article in the series, which lays out how to assemble the ultimate public startup knowledge base.