Gabriel Weinberg's Blog

On the cusp of something big

2013-05-08T00:55:54Z

Startups are a long-term game. My best advice is to treat entrepreneurship as a career path, but it is easier said than done absent some amount of success. For me, I had a taste of it three years in and some real success six years in. Then I started all over chasing at windmills and it was another four years before DuckDuckGo looked like it was turning into a real business. Now I'm entering year fourteen.

I get asked a lot what keeps me motivated, especially from the perspective of a solo founder. I never really had a good answer until now.
I would previously say it is has a lot to do with personality, which of course impacts decision making. I still think that's true though it is not much of a useful answer since you don't have much control over your own personality.

Thinking back, a lot of my motivation has to do with a feeling that I'm on the cusp of something big. I pretty much always feel this way. If we just do x, y, z then this is going to be big. There is always something around the corner, within grasp.

For this to work you can't be completely delusional, only partially. The difference is having a plan.

Of course plans change and things don't always work out. That's why picking an ambitious idea is advised; in a huge market almost everywhere you turn you are on the cusp of something big.

Dealing with a real life externality

2013-04-24T13:13:17Z

I live about a half mile away from a nice volunteer fire department. They're great except for the fact that they blast a siren whenever there is an emergency reported. That's OK during the day, but they also do it in the middle of the night and it always wakes me up.

This is not a fire engine siren, but just a general notification siren. It is very loud. It goes on for a long time and wakes me up sufficiently that I have trouble falling back to sleep.

This situation is a negative externality -- a situation where someone is negatively affected by something they aren't involved in. The canonical example is air pollution from manufacturing. The pollution causes health effects to people in surrounding areas who probably aren't involved in buying or selling the manufactured goods.
A few months ago I hit my breaking point. I was woken up three times in one night, and I sent an email offering my help to improve their systems. I figured there must be a better way to notify people, e.g. text messaging, which has a number of benefits (larger range, can give contextual information, can send to on-call people, can have a subscriber model, etc.).

I got a nice reply from the fire chief including this reasoning for the current siren.

We currently have a group text messaging system for emergency and non-emergency notifications to our members, it works well but is imperfect like any technological solution.

The primary reason for the siren still being in use in 2012 is to notify drivers in the area of the station that there is some type of emergency going on and to be on the lookout for emergency apparatus entering the roadway. That being said, we don't have a lot of traffic in the middle of the night that would warrant it. It also does serve as the primary notification device for members who may performing property maintenance and cannot hear their personal paging devices going off over lawn mowers, other small engines, etc. Also, not an issue at night.

We have had an electrician come out to provide a quote to put a timer on the siren that will limit its hours of operation. The bottom line, its just about $500. We (The Company) concur that the siren is not needed in off-peak hours, but have much more important things to budget and pay for then this. We are willing to accept a donation in this amount to pay for the switch, and we would schedule to have the work done as soon as possible.

I read this and immediately paid the $500. I also kicked myself for not making this inquiry years prior.

They did promptly put a switch on the siren and now I am not bothered at night. Negative externality solved. I have no idea how many others were effected by it, but I suspect a lot. If it had been a lot more money I probably would have done some community outreach and taken donations.

I'm generally good about inquiring about things, but this one got away from me. I had asked through my neighborhood to tell the township, which they did, but nothing happened. If I had known it was a relatively small cost earlier, I obviously would have done it earlier.

I wonder how many other small negative externalities are out there waiting to be solved.

Techmemes

2013-04-19T13:24:33Z

I really needed some levity this morning, so we (at DuckDuckGo) matched some tech companies to memes.

If you think of other good ones, tweet them at me.

Constructing a better startup pitch

2013-04-09T00:47:28Z

When I'm thinking about investing in a startup, I first tell my wife about it and give her my version of their pitch. If we do invest, I often find myself doing the same type of pitch to other angel investors.

My pitch often feels very different than the company's pitch. In a previous post I encouraged people to distill their pitch down to a compelling story and get rid of everything else. That's what I do.

After another year and half of angel investing, I have a bit more clarity on how I think you should structure your story. You want to address the following questions in this order.

Why you? I always find myself telling your personal story first. How did you end up doing this, that is, get into startups, this particular idea and with these particular people? Why is your team the right team for what you're doing?

Of course, if you have any traction, it can go here too. Traction shows your team may already be a good fit.

You don't have to have an amazing track record to have a good personal story. Passion and unique insight are powerful. Personally a major part of what is important to me in this section is that your team cares so much about solving this problem that you'll want to stick it out for at least five years.
Why now? Jason Calacanis has a great Pandodaily post on this question.

When Sequoia's partners ask startups this simple question, you see gears and wheels turning furiously behind the eye of these promising founders being sent into the arena. They start reflecting on perhaps one of the most important questions in the age of the insta-startup: What has changed that makes this the perfect time for this startup to exist?

Every great startup has a "Why now?" when you think about.
YouTube's "why now?" -- bandwidth costs plummeted, video cameras started recording to digital files, and these new smartphones were coming online with -- wait for it -- built-in cameras AND an Internet connection.

Answering why now takes your investment opportunity from interesting to compelling. At it's best it shows you know a secret. It should at least mean you've really mapped your space. Why hasn't it been attempted yet? (It probably has many times.)

This section essentially conveys somebody is going to do this soon. And if you answered the first section effectively that it is also clear your team is as likely as any to get it done.
What is your path to victory? This is your execution plan. It shows you're not crazy or quixotic, or whatever you want to call it. You have a plan. It will change, of course, but given all the uncertainty here's what you're planning on today.

The best paths come with numbers of some kind, e.g. a spreadsheet or some back of the envelope calculations. These are not made up financial projections, but a list of clear assumptions. Your plan is to de-risk these assumptions and reduce the error-bars on them.

Essentially this section conveys how you think disruption will unfold in your space, and how you're going to be a part of it.
Why me? The story ends with real, but intended flattery. Why are you coming to this particular investor at this particular time? How do you imagine they will add value?

The traditional pitch advice tells you to address all the risks investors evaluate, usually with one or two slides each. Better advice tells you to try combine some of those into a six-slide progression, which is what I did when raising money.

The story pitch furthers that progression -- still only a few slides, but weaved together a bit more persuasively. It should create more of an emotional connection than just purely an intellectual one. The Why you piece helps people get to know you. The Why now section adds a bit of fear of missing out. And Why me is again personal.

If you try to use this format, I'd love to hear about it.

Apply to Open Angel Forum V

2013-03-25T13:14:19Z

Applications for the next Open Angel Forum in Philadelphia are now open! If you are involved in a startup that is looking for seed funding, please apply (there is no charge to apply or attend if selected):

Application URL: http://ye.gg/oaf

Application deadline: Sun April 14, 11:59PM ET
Event date: Wednesday, May 8th (private event)
Event location: First Round Capital
Requirements: US tech startup with at least a demo

The idea behind Open Angel Forum is to get the best regional angels in a room and put some great tech startups in front of them. Antonio Tedesco and I started this Philly area chapter in 2011, and this will be our fifth event.

Many of the previous startups who attended received funding thereafter, and we expect a lot of the same great investors to attend again. We've also been working to pull in other active independent investors in the region as well as newly minted angels.

If you are an active, independent and tech angel investor in the mid-atlantic region and want to come, please let me know. We're also looking for additional sponsors since the event is free to both the startups and the investors.

Finally, if you are interested in the Philly startup scene, come check out the first mentor chat I'm hosting next week.

Which investors will invest in your startup?

2013-03-20T18:25:22Z

I get asked a lot by founders which investors might invest in their startups (as I'm sure most investors do).

There are two sides to the answer. The first side is what investors are right for you? I'm not going to cover that side in detail here because it is a big topic and I want to cover the other side, but know that side is just as important. You want people that can help you but also that you really want to work with for many years.

The second side -- which investors will actually invest in you -- depends on your traction. I made the above graphic to illustrate this point.

Traction is of course in the eye of the beholder (and that's one problem with using the word itself). Generally speaking though, the more sustainable growth of engaged users/customers you have, and eventually the more sustainable growth of revenue/earnings you have, the more investors will be willing to invest in you.

The smaller your perceived traction the more you have to focus on investors that would invest in your idea and team. OK, so who is that?

When you're just starting out and have just an idea on a napkin, you're really asking people to invest in your team. There is so much risk in the execution of your idea (which may likely change dramatically) that investors are more making a bet on you personally.

There are currently two groups of investors that regularly invest only in teams:

1) Friends and family; in other words, people that really know you. I'm surprised how little I see founders take friends and family money (including their own money). It actually sends a very positive signal to the next series of investors even if it is a smallish sum of money, say 20K. You had enough conviction to put in your own money and put your reputation on the line with your family.

2) Accelerators. There is one or more in almost every major city now. This relatively new funding source dramatically opened up early money to teams, but you generally can't get too far on it alone.

A third category is also emerging in crowd funding.

After you get your team moving and have started actually building something (but don't have any real traction yet), you open up the investor universe just a little bit further to include investors who really buy into your idea. Who is that? Those are people who generally have deep expertise into what you're trying to disrupt. Look for investors that invested in the same general technology area (e.g. search) or previously sold a company similar to yours.

It is very hard to to convince investors when you have no traction. Your story itself is either going to click or it isn't, and it is much much more likely to click if the investor deeply knows the background to what you're doing already. Think about it. If the investor doesn't have the knowledge to see the obvious disruption that you see then they'll have to educate themselves on all that background knowledge to get to the same point you are.

Once you have some traction, the universe opens up further. But check out the chart -- even with a lot of traction the universe of investors that will actually invest in you is still small.

Investors have investing theses -- these can include a whole host of factors like geography, market size, technology area, valuation, amount invested, ownership requirements, control requirements, etc. When you match your company to these investing theses, most investors get excluded even for companies that have a decent amount of traction.

Luckily the investing theses are pretty well-known. A quick proxy is to look at companies an investor has already invested in; that is, their portfolio. If these seem to match you then you're probably in the right area.

Another way to look at this whole question is risk. When you start out you have a lot of risks in different areas, e.g. team, technical, market, financing, etc.

As you de-risk these areas with actual traction, more investors will want to invest in you simply because the deal is less risky. You also get valuation bumps as a result.

Your friends and family don't need traction to de-risk the team because they already know you. Knowledgeable investors (in your specific area) need less traction to de-risk technical and market risk because they already know your area.

As a corollary, if you're out raising money and people think you're crazy, you don't have enough traction yet to be talking to those people.

The main takeaway is to make sure you are focusing your efforts on raising money from the investors that might actually invest in you.

Recent tasks I sent to my virtual personal assistants

2013-03-12T19:08:40Z

I find myself recommending virtual personal assistants to people quite frequently (among other services). I personally use Fancy Hands. (No, they did not ask me to write this post.)

I don't think many people I've recommended it to have converted, however. It's a way of life change, and those are hard to make.

"What do use it for?" is the standard question. To answer that question more effectively, here are my last twenty tasks (abridged):

Find me a high-quality black yoga mat on Amazon Prime.
Find me replacement hardware for this Restoration Hardware table.
Schedule an oven repair given these error codes.
Make dinner reservations tonight.
Schedule our regular painter to come out and give us an estimate.
Schedule our car to go in for a recall and mirror replacement.
Find out why our corporate filing never went through.
Find the gross profit margin, revenue and # of users for a variety of tech companies.
Buy us a membership at a local kids museum.
Sign us up for sewer insurance.
Help me find an interview of Ira Glass interviewing Terry Gross.
Get a kick-plate for our new fridge, which was missing from the packaging.
Research co-working spaces in Brooklyn given some parameters.
Figure out how to replace the glass on a phone cheaply.
Figure out how to replace a custom wood vent in our house.
Schedule our other car to go in for a battery and radio fix.
Find a place to donate some specific furniture to.
Schedule our regular electrician to come out.
Fix a messed up order with Verizon FIOS.
Schedule a shower door company to come out and give us an estimate.

These are all from the last six weeks. We just remodeled our basement so there is a bit more house stuff, but generally I use it mostly for:

Scheduling (reservations, estimates, etc.).
Web research (purchases, lazyweb, etc.).
Making any phone calls possible I don't have to make (orders, support, etc.).
Arranging travel (booking, what to do, etc.).

Fancy Hands also has a common requests page.

If you're on the fence, I suggest buying an unlimited plan somewhere for a year. That will really commit you up-front and force you more than otherwise to make that way-of-life change.

Reduce communication overhead through more dynamic channel choice

2013-03-08T18:52:45Z

I routinely use these communication channels when discussing business stuff:

In-person meeting
Audio (e.g. phone)
Video (e.g. Skype)
Personal email
Mailing list
Instant message (e.g. HipChat)
Text message
Online task manager (e.g. Asana)
Public forum (e.g. Github issues)

Using the best channel for a given situation can significantly reduce communication overhead, to the point where channel choice deserves conscious thought each time. It sounds obvious but I see people making the wrong choices all the time. Sometimes it is a selfish choice. Sometimes it was a well-reasoned, but wrong guess. Sometimes it is inertia because that's the way it has been done in the past.
I try to make an educated guess about which channel to use based upon the objective and what I know about how the other people involved use the different channels. Some people simply don't respond to email in a reasonable timeframe, for example. They may not check it often, may not organize it well, may feel they can ignore it, may get writers block frequently -- it doesn't really matter why -- they are simply less responsive there.

The second dynamic choice is to know when to jump channels. If I notice a thread stalling, getting out of hand, or giving other signs of unnecessary communication overhead, I think if we should switch to another channel.

For example, a long threaded asynchronous discussion might need to be upgraded to a more synchronous discussion to get something decided in a desirable time-frame. And in that upgrade situation you further have the choice of various levels of bandwidth (text, audio, video, in-person). These are tradeoffs in upfront time (e.g. scheduling) that can be made up in the end many times over by ensuring everyone has bought into that decision appropriately by the time the meeting is done.

It's more of an art than a science because it involves individual personalities.

The depressing math behind consumer-facing apps

2013-02-22T17:22:22Z

You just hit your millionth active user -- congratulations! Unfortunately, getting to that next order of magnitude (10M) is going to be very difficult, costly or both. Unless of course you're inherently viral. There is a reason Twitter, Facebook, Instagram, Pinterest, Dropbox, Snapchat, etc. all are. It's the same reason USV has their investing thesis.

If you're not growing significantly organically, you have to use a traction channel and either go out and reach people (ads, press, etc.) or set something up so that they are coming to you (SEO, content, etc.). Either way you will have to convert them.

Conversion ratios for viral are awesome. Conversion ratios outside of viral are not.
In a mass-market setting, decent conversion ratios are 1-3% depending on vertical, 5% is amazing and 10% is essentially unheard of (please give counter examples in the comments if you have any). For your first million, you could be somewhat niche and get higher ratios in smaller-scale channels, but when you're going up to 10M you are really crossing the chasm and sadly reverting to the mean.

Let's say you're really good: 5%. You need 9M new users, divided by 5% is 180M highly-qualified impressions (math gets a bit complicated when you convert that to people and start accounting for overlapping campaigns and conversion ratios increasing with # of individually experienced impressions).

That is a lot of impressions. For example, an amazing NYT feature won't even get you close (unless it spawns a fire-storm of other mass media press). Their circulation is 2M on Sundays, higher if you count online, but back to that if you account for people actually reading your amazing feature (remember, it's highly qualified impressions).

You can begin to understand why AOL printed all those CDs. And if you have high churn, god help you. It's a little better if you have some monetization so at least you can afford all those CDs.

Otherwise, you better raise a boat load of money, if you can. It's an easy sell to raise a ton if you haven't figured out monetization but already crossed the chasm (Tumblr). If you haven't though, it's doubly risky with both high market and finance risk. (Yes, this my obligatory paragraph required for all startup articles in 2013 to relate in some way to the Series A crunch.)

Some sectors are better than others. If you have low churn and high monetization (e.g. finance) you have more options like being in the top ten advertisers on the Web. If you have high churn or thin margins, however, it gets tough as Fred Destin pointed out for e-commerce. In his words, what's your angle? Or if you are raising money, why now?

Either way, never count out organic growth in your traction plan. If your first million active users are truly actively engaged and have pretty wide demographic characteristics, there is hope for you after all. Cut through this depressing math with the community sword and bump up that conversion rate!

Update: some good comments on HN.

Some ideas matter, just not the ones you think

2013-02-10T13:48:30Z

Startup wisdom holds that startup ideas don't matter too much because execution is everything. That's largely true, though I think markets do matter and the line between markets and ideas can sometimes be unclear. Also there are plenty of edge cases of ideas that do matter (powder-keg ideas, a deeply knowledgeable team that knows how disruption will unfold in 1-3 years, etc.). I'm not exploring these distinctions here though because I find this debate largely irrelevant in practice. Unless you have a track record, you generally need some traction anyway to get funding; and if you don't need funding, just go execute.

Within execution though, there are ideas that matter lurking. It is those ideas I want to unpack a little. If you have a product development and traction plan, on the surface you can break that plan down into smaller and smaller pieces and then execute them. I'm going to build this and that feature; I'm going to test Facebook ads; I'm going to try to get promoted on x, y or z.

But when you get into it, there are many ways to synthesize product feedback to decide which features to build and how to build them. Similarly, there are many ways you can present your startup messaging to potential customers and the media, and many traction channels to pursue.

In that space of sub-ideas lurk ones that resonate way more than others, like resonant frequencies (see image). Sometimes these are as simple as a few new words of copy or a slight UX tweak. They are essentially powder-keg ideas within your startup, waiting to be sparked and move the needle.

These ideas matter, and you should be designing processes to surface them as regularly as possible. For each startup team these processes will be different, but here are some general things that I've seen work for me and others.

First, some obvious ones:

Have a structured process for dealing with product feedback. Don't act on everything, especially immediately. But log it, synthesize and try to understand why it is occurring.
Have a structured process for getting traction. I advocate our Bullseye Framework, but really anything is better than nothing.
Always be experimenting. It doesn't have to be explicit A/B testing, though that is good in many situations. But more generally, if you're not experimenting then you're largely decreasing the probability of uncovering resonant ideas.

And some less obvious ones:

Horizontally integrate your team, i.e. have them work on multiple areas of the business. For initial founders, I advocate both founders work on product and traction equally. The basic premise here is two-fold. First, outsiders to things coming in often come up with interesting ideas. Second, if you're siloed then you can't easily come up with holistic ideas; oppositely, if more people are exposed to feedback, marketing, product, etc. then they are more likely to do so.
Consistently put yourself in a position to be more creative. For me, it is not operating at capacity; for others, the opposite. Lifehacker consistently has good tips like lying down.
Do everything you can to be empathetic with your users. Be one if you can, but failing that, watch new users, talk to power users, pester potential users about their problems and how they perceive solutions. This customer development is pretty standard practice up front, but seemingly often gets left behind when product development starts in earnest.
Find a consistent place to test messaging, e.g. reddit ads, AdWords -- something with a potential audience that you can tinker with.

I can't impress enough that these resonant ideas are often very small. We've seen tiny wording changes change CTRs from 1% to 5%, for example. When I put up our billboard we doubled our user base that month.

You can think of these ideas as regular usability or marketing improvements, but I think the most resonant ideas are much more. The truly resonant ones change the perception of your product. They click with peoples' previous conceptions of the world. And that is exactly why they are so powerful.

They are psychological. They are the ideas that help you put your foresight of how disruption will unfold into useful practice. And that is why they matter.

Resonant ideas matter in the same way people argue about startup ideas mattering: they are inherently valuable. However, they are nevertheless hard to copy because they usually involve a code iceberg of some kind. You do a small tweak on top of your existing product and marketing infrastructure. The infrastructure is needed to succeed, but you still need to hit the resonant frequency to make the impact you want.

Update: some useful comments on HN.

How will disruption unfold in 1-3 years?

2013-02-09T21:25:07Z

I remember someone asking me six or seven years ago what was next for mobile phones. I said something to the effect of everyone's going to have a smart phone and be able to watch video on it. Prescient? Not usefully. It's like predicting the demise of newspapers in the mid-nineties or long-distance in the early-nineties.
It is relatively straightforward to carry certain trends to their logical conclusion and come up with plausible predictions for some point in the future. In the long-run everything is disrupted. But it is much harder to figure out who may win or lose from that disruption and what new business models or tweaks will rise up. It is harder still to put actionable timelines on those predictions.

In other words, predicting general disruption doesn't put you in a position to profit from your prediction. To do that, you need to answer a much more specific question: how will disruption unfold in 1-3 years?

If you can answer that question you can create a path to victory and stay on the right side of crazy. If you can't, then you should probably stay away.

Unfolding significant disruption in under a year is possible, but extremely rare. Those are the powder-keg ideas that don't take a lot of development time (arguably snapchat most recently).

Having a long-term vision (say 5yr+) is also important, but that's too far in the future to be actionable (at least in Internet tech). You should be able to work backwards from that vision and create a more concrete path you can execute on in the short (6-18mo) and medium (12-36mo) term.

This path also helps you communicate for fundraising. The story you're telling investors should be about this path. Savvy investors understand the difference between actionable and unactionable disruption predictions. And traction further communicates that you're right.

The Bullseye Framework

2013-01-22T13:30:00Z

Not using any structured approach for getting traction is one of the most common traction mistakes, and yet there are simply not many traction frameworks out there to help you make sense of the process. In our upcoming Traction Book, Justin and I present the Bullseye Framework for getting traction. It is a simple approach that is meant to act on the following premises:

In any given growth phase, usually one customer acquisition channel (traction vertical) is largely responsible for really moving the needle for a startup. This is a consequence of the fact that getting traction itself is usually (though not always) a power law problem.
There are about 20 different traction verticals that we see startups use to get traction. It is hard to predict beforehand exactly which traction vertical will work best for your startup at a given time. You can make educated guesses (and you should!), but it is hard to tell until you start running tests whether a given vertical is the best one for you right now.
Most startup founders have a significant bias towards (or against) using certain traction verticals, e.g. for ones they're already familiar with and against ones they know little about, find icky or involve a lot of perceived shlep. A corollary of this premise is that there are far too many startups focusing on the same traction verticals and that many potentially promising verticals get overlooked and are therefore under-utilized.

Our framework for approaching traction uses a bullseye metaphor (see picture). At any given time, the inner circle contains your most promising traction verticals; the next concentric circle contains the next most promising, and the latter rung contains the rest.

Using the Bullseye Framework is a five-step process:

Step 1: Brainstorming. Go through each traction vertical and brainstorm how you might use it effectively for your current situation. This first step is meant to help you systematically counteract your biases by forcing you to take each vertical seriously. As such, it should take a while. First, you need to have a basic understanding and appreciation for each vertical, which our book is designed to give you. Second, you should have mapped your space to know where others have succeeded and failed at each vertical and why. Third, for each vertical you should spend at least an hour thinking up plausible paths to traction within it, and more specifically possible cheap tests to run in that vertical. The whole process should be spread out over days or weeks.

Step 2: Ranking. Place the traction verticals into three columns:

Column A (Inner Circle): what traction verticals seem most promising right now?
Column B (Promising): which traction verticals seem like they could possibly work?
Column C (Long-shot): which traction verticals seem like long-shots?

Step 3: Prioritizing. Go back through your columns and critically re-think your choices. Try to place your top three traction verticals in column A, the next six in column B, and the rest in column C.

Step 4: Testing. Your top three traction verticals in Column A comprise your inner circle. Devise cheap tests to experiment with them in parallel.

Step 5: Focusing. If one traction vertical from your inner circle shows significant signs of early traction, then you have your winner (for now). Double down on what's working, and wring as much traction as you can from it. If none of the three traction verticals from your inner circle proved successful enough to focus on, go back to Step 1 and start over, incorporating your new-found knowledge (from the other steps).

Why three? Why test in parallel? There are a bunch of nuances in here that we expand upon in the book, but the basic Bullseye framework is pretty straightforward.

We argue you should spend about half your time pursuing traction (as not spending enough time is another common mistake). Whether you use our framework or another, it is strongly advisable to take a structured approach to getting traction.

Go talk to founders who failed at what you're doing

2013-01-17T16:50:07Z

It bothers me when someone tells me about their startup and I ask how it relates to xyz company that did something similar in the past, and they have no idea xyz even existed. It's a negative signal common to first-timers.

I realize there is tremendous value in having a fresh perspective. But in startups there are so many paths to failure. If you are going to be a successful heat seeking missile, it really helps to know what has happened in your space and related spaces.

At worst, it will help you avoid mistakes and stay on the right side of crazy. At best, you will uncover secrets that can help you get traction faster.

The best way to map your space is to go talk to the founders of xyz company. They failed at what you're trying to do and generally they are very willing to talk you about it. In fact, you may be one of the only people in the world that will truly get what they spent all that time doing.

It's an outlet for them for pent up information and in some cases latent emotions. For you there is a lot of upside: a quick history lesson and potentially extremely valuable information. And among these conversations you may even find a mentor or investor.

Thank you David Horowitz for reminding me last night of this great tactic.

Orders of magnitude

2013-01-06T13:38:53Z

When people say they're looking to grow by an order of magnitude they usually mean ten times.[1] For example, if I grew my visitors by two orders of magnitude that would be 10x10 = 100 times.

I find framing things in orders of magnitude is a really useful way to measure progress and think about the future. Not much changes structurally if you grow by a factor of two; usually your technical and non-technical infrastructure can handle that kind of growth pretty easily. But when you grow by a factor of ten (an order of magnitude) something usually breaks. And yet thinking much beyond ten times can be very challenging because you'd look a lot different then (as many things would break).
An order of magnitude of growth also makes you a substantively different company. Generally that would mean you have ten times more customers. At that level it can be harder to move the needle with the same traction verticals, or will at least require significant changes to do so.

Since starting, DuckDuckGo has grown by about 4.5 orders of magnitude for the one metric that matters to us, real direct searches. We were at about 1,000/month after launch settled down and now we're at about 50 million/month. I'm really proud of this growth, but we're still about one order of magnitude from making a measurable dent in the search market and two from a major one. You can easily put yourself in perspective by thinking in these terms.

Needing that many orders of magnitude to make a dent speaks to the size of the search market, and more generally to consumer Internet. There are billions of people on the Internet now, and the most mass-market services are going after at least 100M of them, and the very biggest all over them. Non-mass-market services (e.g. B2B) operate in fewer orders of magnitude for customers, though the biggest reach similar amounts in terms of revenue, e.g. $10,000/year to $1,000,000,000 -- that's just five!

While things may change internally when you grow by an order of magnitude, to the outside world things may not look that different. To many people who weren't using our product we probably seemed the same during these order of magnitude changes and still do. Even to savvy investors I speculated our valuation may not have changed much through periods of that growth, and that makes some sense. After all, even though you may have to operate differently to absorb and create another order of magnitude doesn't necessitate that your expected terminal value will change significantly.

[1] Technically, it doesn't have to be base 10, but 10 is assumed if not otherwise specified. I got to thinking in order of magnitudes in college doing my Physics degree where it is common to ask yourself "do I have the right order of magnitude" for this or that answer, or even more colloquially to tell someone "it is on the order of..." In Physics, you deal with very big and very small numbers (sometimes in close proximity) and so it ends up being an effective way to communicate, at least it was for us lowly undergrads!

Being on the right side of crazy

2012-12-23T18:07:39Z

It's common startup wisdom that you need to be relentlessly resourceful in the pursuit of growth, and doing so may take many years and lots of schlep before success at scale. What is less well understood is how you go about systematically pursuing traction and how you determine product strategy (as opposed to product development). This post touches generally on this latter topic.

One of the best things I've read this year is Rob Go's Path to Victory post, which finally entitled something I've been asking entrepreneurs (and myself) for a long time: what does success look like in the next few years if all goes well, and what needs to happen for that future to unfold? That is, what is your path to victory?

It's really a short-form and focused business plan. As an example check out Chris Dixon's plan for SiteAdvisor. While it's true business plans aren't generally read by investors, a succinct explanation of your path to victory can secure financing quickly. It takes the form of a story that can be put in a pitch deck.

Your path to victory should be on the right side of crazy.
There is some crazy in there. You want to be ambitious enough so when you first start out and you tell people what you're doing the initial reaction of most people is that you're a bit crazy for taking it on in the first place. Then you know you might be onto something big in the same way good ideas are often dismissed as toys.

The wrong side of crazy is when the probability of success is too low for your risk appetite (and for investors if you want and seek them). Laying out your path to victory helps assess the magnitude of that probability.

A common issue is believing that something plausible is true. It's not. Yes, plausible things can create good stories, but underlying them are one or more assumptions that have associated probabilities of occurring within a given time frame. It is your job when constructing your path to unpack those probabilities in the context of your particular time frame.

I don't want to overstate the quantitative nature of this exercise. I think you should have a spreadsheet with your plan assumptions, but at the same time everyone (including yourself) should recognize your quantitative assumptions probably have huge error bars. In fact, you can plan for that in your spreadsheet with bad/good scenarios and more systematically with sensitivity analysis and scenario planning.

A related issue is a path that relies on larger market movements, especially mass changes in consumer behavior. Such changes can create empires, but they can also be hard to time if they haven't started snowballing yet in earnest. Countless entrepreneurs (including myself) have been too early to products by many years. You can combat this problem by being a world-class expert in the changes you think are about to occur, but it is very hard to not have cognitive dissonance when doing so. It is much easier to pursue an existing large market, or one that has already started trending.

Another issue is what Josh Kopelman called the Domino Rally Business Model, where success relies on a series of independent things happening (dominoes falling) in your favor. The problem is since they are all independent events and you have to have all of them to be successful, your overall success probability is the multiplication of the probabilities of each of the dominoes. So if there are four things that need to happen and each are 50/50 you are now at a success probability of (1/2)^4 or about 6%. The chicken and egg problem is a subset of this issue.

Perhaps the best way to construct a useful path to victory is to put more focus on pursuing traction early on. The better you understand your customers (the first step being to get some), the easier you can articulate how many of them there are, how to reach them and how to monetize. Those are in essence the underlying questions you're answering in a path to victory (although you may choose to put off monetization in that path).

Being on the right side of crazy is having a path to victory that is probable and yet ambitious enough to have maximal impact when you succeed. Great paths allow you to incrementally make progress on them and build to inflection points where you get step function growth. Worse paths require large periods of time with no progress until you hit those inflection points (ex. the Domino Rally).

The funny thing is when you take a step back the growth curves in the success case of both these paths look similar. That's because inflection points make s-curves and the difference between slow growth and flat growth before and after the step ups are not very distinguishable on first glance.

But there is a big difference in there. Things that are growing (albeit slowly) have data flowing back to you. This data makes it easier to tweak your path to victory to ensure you actually get to victory.