September 2010 Archives

Yes, PHL/BOS/CHI/BOU/SEA/NYC/BLT/XYZ does Web startups



I have nothing against SV, and I'm not predicting its demise. But I also know that I don't live there, probably won't ever, and don't like to travel.

In other words, I have a vested interest in making the startup scene close to me as good as it can get. I'm pretty tireless in that effort, and lucky for me, I'm not alone.

Yes, Philly does Web startups too. In the past few years pretty much all the pieces of the puzzle have fallen into place, e.g. hackers + mentors + accelerators + money + community = exits. From watching it grow, I think it's pretty clear this can happen in any decent sized city at this point.

Nevertheless, why do a startup in Philly (or any of these cities) if you can move to SV? From my perspective, the only real difference seems to me getting the best investors, i.e. those with the best domain knowledge. (Side note: I'm not saying investors are the be all and end all anyway. My last company had no investors, and so far neither does my current one.)

Yet that final difference is crumbling rapidly with AngelList, Open Angel Forum and other market shrinking efforts. In fact, the "right" investors are now actively seeking you out in these other cities. They realize things are cropping up out here and don't want to miss out on the next big thing. Also, our valuations are more reasonable, just like our house prices.

And then there are the benefits:

  • You're already here and you like it. Quality of life. You may be like me--have a family and no plans on moving. Or maybe you went to school here. Or maybe you joined an accelerator for a few months. Whatever the reason, given that we have everything you need here, there is no immediate pain point forcing you out. So then quality of life becomes a benefit.

  • Easier bootstrapping. Cost of living is just lower, at least in Philly. Also, it seems really easy to pick up extra consulting hours to pay the bills. I did that in Boston for several years before my last startup took off.

  • The Web "stars" are more accessible. The community is smaller, which means people usually go to the same events and are just generally more approachable. Additionally, since everyone wants their local area to have a better startup scene, these people will often go out of their way to help you succeed, even more so than they would otherwise.

  • Easier to get attention. Also as a function of size, if you do something noteworthy, you get noticed. This also works with consulting--it's pretty easy to make a name for yourself very quickly.

  • Easier to hire. Listen to this snippet from This Week in VC about competing with Google/Facebook/etc. for talent in SV. Frankly, that doesn't happen out here.

So what's missing? I think what's missing (at least in Philly) is more ambitious startup founders. The angel investors are here, ready to invest in your company and help you succeed. There is frankly a glut of good investors and mentors. Personally, I think that's a good problem to have and will naturally work itself out.

DuckDuckGo implements HTTPS Everywhere

"HTTPS Everywhere is a Firefox extension produced as a collaboration between The Tor Project and the Electronic Frontier Foundation. It encrypts your communications with a number of major websites."

DuckDuckGo is one of these websites. And now when you use the encrypted version, the same ruleset in HTTPS Everywhere is now applied to DDG links.

For example, that means that all links to Wikipedia will go to their encrypted (https) site. Here's the full list of sites it works for at the moment: Amazon, DDG, EFF, Facebook, Gentoo, Gmail, GMX, Google,, Ixquick,,, Meebo, Microsoft, Mozilla, Noisebridge, NYTimes, PayPal, Scroogle, Tor, Twitter, WashingtonPost, Wikipedia, Wordpress and Zoho.  

You do not need to install the add-on to get this functionality, or even use Firefox. It is built into the back-end of the https version of DuckDuckGo. I've updated the privacy policy to include a mention of this feature.

I thought it was fitting to complete and launch it today because DDG is currently featured on Slashdot for our Tor exit enclave. It's also the 2yr launch anniversary--hence the party duck currently on the homepage.

How-to get that guy as your mentor


Forming a couple of good mentor relationships can help bridge the gap between startup failure and success, especially for first-time entrepreneurs. But how do you actually get the right people to be your mentors? I previously explained that it usually requires some equity, but here is some more step by step practical advice.

  1. Identify people that would make good mentors, i.e. with the right domain experience. This is a brainstorming session, so think big. List out all the right people you can think of and that you find through your market research.

  2. Look through their LinkedIn/Crunchbase/etc and see if they do angel investing/formal advising/sit on boards. If they do, they will be much more likely to be your mentor and more importantly much more likely to be part of a good mentor relationship because they've had experience doing it before. You could go even further in this step by talking to their portfolio companies.

  3. Prioritize your list from #1 given what you learned in #2.

  4. Look for a warm intro to them (LinkedIn connection, friend in one of their portfolio companies, etc.). If you can't find a warm intro, know that some people (including myself and hacker angels) are totally fine with cold intros. It really depends on the person, but warm is always better so I'd really try and find a warm intro first.

  5. Send an email introducing yourself. This is the key part because you're making a first impression. Ultimately, if it works out there will be an equity exchange of some kind. No one wants to take equity in a company that isn't going anywhere, and companies go somewhere because the founders take it there.

    Yes, you're asking for advice, but you're really the one who is going to do almost everything. So you really need to convey from the get-go that you are the kind of person who can make things happen.

    Now I'm sure different people like to see different things, but this is what generally works for me (and I suspect others as well). First sentence say something like "Hi, my name is Kyle XY. I'm a first time entrepreneur starting a company doing Z and could use some advice."

    Then write one paragraph about your history and how you came to be doing what you're doing, e.g. "I went to this school, and then started this in college and then did this and that and I had this idea and I've done X,Y,Z for it."

    Then write one paragraph about the company status, e.g. "I have a prototype, and there are 2 founders and 1 advisor and we're at this and this stage."

    Then a final paragraph asking for particular advice, e.g. "We're not sure what to do about this and think you would have a good perspective on it--what do you think?"

  6. Let's recap. You just wrote a four paragraph email, i.e. relatively short but not too short, to a person who actually does advising, asking for specific advice. Note that you did not ask them to become a formal advisor right now. Why? Simply because they don't know enough about you (and you them), so why would they (or you) jump into such a serious relationship?

    You need to feel it out first, and starting a casual thread is the way to do that. It gives you the option of asking for real advice and seeing if the advice you get back is useful. It gives the other side the option and push to engage in a specific way and see how you react to their advice as well as how you communicate with them.

  7. Over the next weeks/couple months, try to keep the thread going. Don't write back every day, i.e. when it dies down wait a couple weeks and then ask another question, then a few more weeks and another. The assumption here is that you have real questions to ask, i.e. you really do need advice--that is obviously essential or it will not work. Read my previous post for more on that.

  8. Assuming everything is looking good from both sides, i.e. your company is really benefiting from the mentor's advice and they seem engaged, then after a month or two, ask them if they would like to be more of a formal advisor. 

  9. If they say yes, then offer them an equity package on the one hand and their expected responsibilities on the other, and lay out the offer in the email. As I've said previously this needs to be a) enough advice to be useful on your end and b) enough equity to be useful on their end.

    On your end, that usually means (at least initially) something like we're going to continue emailing like we've been doing (every few weeks) or we'd like to have a monthly Skype call for a few hours and would expect review of the issues beforehand.

    On their end, the equity should be enough where it is motivating to stay engaged. In practice, I think this is 2%+. That is significantly higher than normal advisor relationships, and this shouldn't be one of those. This is a deep relationship. It's analogous (and at least I think of it in the same terms) to an angel investment. If you get less than about 2% (depending on company) that can be diluted down pretty easily to almost nothing on subsequent rounds of financing. Remember they're not getting preferred stock for this--usually options.

    In particular, the equity grant should be vesting. I think two years, monthly makes a lot of sense. I think monthly because that's the cycle time of the advice and two years because at that point a lot of things change in a startup company. You can always re-up. The vesting of course allows either side to cut-off the relationship with minimal harm.

  10. Expect a bit of back and forth about the offer. Some points that may be of concern are your financing plans, triggers (what happens if you get quickly acquired, for example), etc. Wrt financing, the presumption is you would offer them a place in your round. 

And you're done! Well, just beginning...

P.S. I have rapportive installed and so see your LinkedIn picture when you email me. So check that out. I also like to look at peoples' blog/twitter etc., so just keep that in mind as well. I suspect others do the same.

Avoid entrepreneur mistakes with good mentors



Recently I wrote about wannabe and first-timer entrepreneur symptoms. I did include cures, but the reality is it is often hard to self-medicate. 

A longer lasting cure is to get a couple of good mentors. The mentors are there to call you on all your bullshit.

You have to be open to being called out, though. In my first companies, I had access to good mentors, but didn't use them correctly and so the mentor relationships were not good (my fault), and the companies ultimately went nowhere. Now I'm not saying they would have gone somewhere if I had done the mentor relationships right, but it certainly couldn't have hurt.

What makes a good mentor? You really want someone who has been there, done that, or in the words of Mark Suster, has domain knowledge

However, just because you've identified people with the right experience, doesn't mean you can create a good mentor relationship with them. That's the hard part.

A good mentor relationship is a series of deep conversations about your startup's underlying assumptions. It is your job (in general) to identify those assumptions and validate them as cheaply as possible. Sometimes though, you can't see the forest through the trees. 

That's where the good mentor comes in. He/she should have a full picture of your business and be able to say, wait, did you think of this OR there is this other underlying assumption you haven't mentioned OR no, that doesn't prove that assumption. 

The core benefit of the mentor relationship is getting someone outside the company to think critically and holistically about the company. Making intros, specific advice, etc. are all good too, but in my opinion they are side benefits. 

From the mentor side, it's a time commitment. It doesn't have to be a ton of time, but at the very least you should be communicating once a month. And these communications should be at a deep level, which requires a free flow of information in between conversations. That means you have to be continually sending them stuff. 

So how do you get these super busy experienced people to do that? There are a few ways, but generally some form of equity compensation is involved. For example, you can get them to do a small investment at a low valuation, or give them some options. 

Note, however, that typical advising shares (e.g. 0.25-0.5%) are usually not enough. Do the math (same math as for employee options). 2%+ depending on their role and expectations is more in line. Shares should vest though, so if it isn't working out there isn't much harm done if either side decides to call it quits.

Ultimately though, the mentor needs to be interested in your particular company. Equity of course helps spike interest, but it isn't enough by itself. So look for people who seem generally passionate about what you're doing. To do that, you really need to start forming a relationship before you ask for the deeper mentor relationship. In the best case, it happens naturally.

Update: check out the follow-up post on getting mentors.

4 out of 5 hacker angels invest in Locately

I'm proud to announce the first company in which a majority of hacker angels invested: Locately. HA is not a fund--just a group of hackers who also do individual angel investing. This means we each individually decide to invest, in this case 4 out of 5 of us.

Locately is leading the location analytics market. They analyze GPS data from opted-in consumer mobile phones to help businesses better understand their customers. For example, they can help companies determine who their real competitors are, if physical advertising is actually working, or where to put their next stores.

There are really a lot of possibilities. Just as web analytics is an essential tool for online marketing, I believe location analytics will become essential for offline marketing.

The easiest way to get investment quickly is to have traction because traction trumps everything. And Locately has traction. Even so, many angels (including myself) want to actually know the founders beforehand as much as possible. This desire stems from the fact that doing a startup involves a lot of meandering and critical thinking, and over time you can start to get a sense of how people think about and react to things, as well as see progress.

So I'm fairly certain that starting dialogs with potential investors long before you are raising money will increase your chances of actually raising money. There are many ways to do it, e.g. sending periodic newsletters, asking for feedback on product/strategy/fundraising/etc., or even bringing people on board as formal advisors. In this case, I communicated with Drew (one of Locately's founders) on twitter, and then we hung out for most of Angel Boot Camp. That really kicked things off.

Over the past few months (since announcing hacker angels), we've dispensed a lot of (I hope helpful) advice. To get some yourself, just email us:

Sean Murphy on the first dozen enterprise customers

I recently did a Traction Book interview with Sean Murphy who runs a boutique cutomer development firm in Silicon Valley. We discussed (for ~1hr) how to approach getting the first set of enterprise customers. What follows is a slightly edited transcript of the full interview. 

yegg: Why don't you start by introducing yourself?
skmurphy:  I'm Sean Murphy. My firm is SKMurphy.  There are six of us that work with early-stage software and technology firms, helping them find early customers and early revenue.  We've been big fans of Steve Blank's Customer Development Methodology for several years now. Our primary focus is on firms that sell to business across a variety of technologies, almost all of them software of some sort.  

yegg: So all enterprise?
skmurphy: Yes. Sometimes we are selling to small/medium, but it's all business-oriented. Only very rarely do we help with consumer or advertising-driven model, and in the advertising-driven it's where that looks like enterprise.  

yegg: What is your role exactly, i.e. when do you typically get involved and what are your typical responsibilities?
skmurphy: So, we get called in a variety of places. Sometimes folks that are trying to come together will contact us and we'll help them with formation.  We'll refer them to attorneys, accountants, naming experts...  Often they are underway: they've made a couple of sales but they didn't realize why they were successful or what they are not doing.  Sometimes they are just pretty close to giving up and they are trying to figure out should they go on or not.  

It's normally early market issues: they typically have fewer than six paying customers. We have worked with some larger firms that are doing new product introduction where they have a well-established product but they've "lost the recipe" for introducing a new product. And we've worked in a couple of turnaround situations where a new CEO has come in who is trying to figure out how to take with the technology in hand and find a market for it.

In most cases our model is the "sell what you have."  

yegg: OK, so let's walk through an example. How do you help a team get that first customer? 
skmurphy: Most of the time it's going to be somebody that either they know or that knows somebody they know, or somebody that we know or knows somebody that we know.  For the most part our clients are going into a market that they understand with technology that they have developed. We help them make a list of every project they've worked on and everyone they've worked with. They reach out and say, "Here is what we are doing, do you know somebody we should talk to that makes sense? "

People who've developed expertise by working in a field for a while are typically able to get an initial meeting--cup of coffee or lunch, these kinds of things.  Sometimes we encourage them to shift to a different market because we find out that the technology has more applicability and offers more value there. 

Let me try and zoom in just a little bit more:  one of the first things we help them with is to put together what we call a lunch pitch. This is a single piece of paper that has five to ten bullets and a perhaps a visual that helps them focus the conversation making sure they understand the prospect's problem.  The early conversations are all about exploring the prospect's problem and pain points.  

yegg: So it is pretty rare that the first customer is from a cold lead?
skmurphy: I haven't seen it. The challenge with a first customer for new software is that big unknown in the prospect's mind is if the team going to perform.

One of the best ways to attack that is through testimonials, but initially it's got to be past performance, where somebody can vouch for the person--you know I've worked with Gabriel at this company or I've worked with so and so here.  

yegg: You sit down with the warm lead for lunch. Is there anything productized at this point or are you in an open-ended discussion?
skmurphy: The first and most important thing is to figure out: do they have a problem that we can help them with?  Often that will initially look more like a mix of consulting and product operation.  So, the technology has to at least work when it's in the founding team's hands.  It's nice if it can work if you give it to the customer, but the founders have to at least be able to make it work to solve a real problem for the customer that the customer will pay for.  

yegg: You have no problem with a consulting aspect for the first few customers?
skmurphy: No, and I think other people have described that in different terms.  There is this term Flintstoning, or Wizard of Oz--it's not uncommon that the team has to either work around problems in the technology or on-the-fly improvise adaptations that look more like a service engagement.  If you try and make your product perfect before you engage, you'll run out of time.  

yegg: Because you didn't understand the problem?
skmurphy: Right. Another thing that can happen is entrepreneurs show us a datasheet or a feature list and they've got forty features listed.  Maybe twenty of them are real and twenty of them are things that they are planning, and they want to go to a customer and say, what do you think of these forty features?  

The reality is that if the first two, three, four, maybe five are not compelling, he is not listening to your next thirty.  So, another thing we'll help startups with is to focus on a core need or a core problem leveraging a core capability. 

We help them focus on the delta between the prospect's status quo and what their offering does to move them to a better level of profitability, level of revenue, lower error rate, faster cycle time, .... You need to proceed from some theory about a delta, some outcome you are going to create in the prospect's business.  

yegg: Let's suppose the prospect is excited. How do you close the sale?
skmurphy: This is a multi-step process. If you are learning more about the prospect's needs you are making progress. A few times in the middle of one of these conversations I have had founders, look at me and say "Well, they are not going to buy, we should leave." It requires patience; software is the promise of relationship. 

Typically, the next thing you ask for is prospect data or some way that you can demonstrate a result which is beyond a canned demo, a demo that uses data from the customer's environment.  

The normal test we tell the founders we are going to have to pass is--after a coffee break or a couple of lunch conversations---that we are going to go to somebody's office, and in less than two hours we have to show them something that's relevant to their business that makes a difference.  And, if you can't do it in two hours then you have to leave with data they have given you and come back within five days with something relevant.

We've seen people take different approaches, but in general if you are on site more than a relatively short amount of time, you are forcing the guy to spend political capital either to get your badged in or to get your access. You want to hold off on that until you've unambiguously demonstrated value. 

yegg: At this point, have you introduced price or any other terms?
skmurphy: Price is pretty slippery in the early market; we are really trying to understand what the impact is on the prospect's business or on a particular business process.  

If business people are spending time for you they've got a real problem that they are looking for you to solve.  There is typically some real pain point that's underneath.  

Now, trying to figure how to value that, and what's needed to solve it, is obviously many conversations and a fair amount of analysis.  

It's a complex thing.  Normally we let the customer bring up price. We will ask them how they evaluate the value we are creating for them.  I know there is the famous Steve Blank question "Would you pay a million dollars?" which I think is more rhetorical than literal. It can be risky in the beginning if you name a number that's too high: instead of countering with their value number, they may stand up and leave the coffee table or otherwise end the conversation.  

yegg: You ask them to give you a number?
skmurphy: Yeah, and we'll tear that apart a bunch of ways.  We've been asking a lot of questions all along in terms of what's the impact on the business.  We are trying to draw a kind of a before and after map either of a process or an outcome, and then using judgment and talking to many people to try and understand the value that's being created here.  

As a rule of thumb you can only capture a fraction of the value that you are creating.  You can't in the abstract just say "This is worth a $100,000" or "This is worth $5,000" or whatever.  

yegg: Doesn't this process typically get low-ball numbers thrown at you?
skmurphy: Sure, and in the beginning, the other thing to understand is that the people that you working with, when they are spending time with you, you have to count that time as part of what they are paying for the product.  

In our experience it's difficult to get the full dollar value out of the first couple of customers.  That's of less concern than being able to build a real case that you are creating value as a result of actually getting them to implement.  They are going to invest typically a fair amount of time on the part of people with real opportunity cost to get something working.  

yegg: So you are OK with a lower price for the first few customers, but you do want some money changing hands?
skmurphy: Yes, yes we do,  typically a nominal amount that doesn't trigger the need for budget escalation or other kinds of negotiations.  And, we are talking about the very first few customers.  We are more looking for success stories.

Where we start to negotiate much harder on price is where they say you can't talk about this, we are not going to give a testimonial, and we are not going to act as a reference.  The payment we are really looking for on the first couple has more to do with the ability to do test cases, testimonials, and work with them as a partner for at least a year or two to really understand the impact of what we've got.  

yegg: Why insist on price at all then?
skmurphy: If money doesn't change hands, you can't call them a customer.  Now, money changes hands maybe four steps or five steps in.  You are trying to first get their attention, their time, get their feedback, and get data. You are asking for them to provide you an evaluation or a benchmark, maybe they do a pilot project--at that point then we start talking about money and testimonials.  This is once they've actually seen that operate in their environment, which looks very different from a standard sales process.  

yegg: This only applies for the first few customers?
skmurphy: Right, for the first half a dozen or so. And there can be a lot of work to find them. We work with a team that has a very innovative technology, and we must have talked to twenty people.  It was a chip design application, and a bunch of them looked at us and said, you know this seems really interesting, but we have no use for this whatsoever.  But the founders had come out of a chip design background and were confident many firms had the problem and more would over time. Sometimes it takes two or three dozen conversations to even figure out who you should really talk to and who really has the problem.  

yegg: How low is too low for the price?
skmurphy: That depends on the company. In general, the lowest level of signing authority is typically $500 or $1,000.  

yegg: You'd be comfortable going that low?
skmurphy: Oh absolutely. For a first customer who will use a new technology in production and give us ongoing feedback we have accepted $1,000 for a first license.

yegg: And these customers are also acting as references for you?
skmurphy: Yes, and we negotiate that much more carefully.  The money is important in the sense that it's not zero because that means it goes through a purchasing process, and it's unambiguous that we can call them a customer, as opposed to my friend at XYZ Company evaluated it--that's not that useful.  

yegg: How do you go about negotiating the terms of the testimonial piece?
skmurphy: We'll offer them very explicit options in writing, not a contract, but a kind of a set of talking points or an English language description of would you feel comfortable taking at least one call a month, would you be willing to allow us to write a blog post about this, would you be willing to let us release a press release about this?  There are about seven levels or eight levels of endorsement that you can seek, and you find the level that they are able and willing to do. 

A lot of times the problem is in larger companies, the people you are talking to are constrained in what they can do publicly.  They may say, for example, I am happy to talk to three people a month, but I can't do anything in public. And that's typically workable.  Sometimes you can get them to do a technical paper at a conference, or you can go to a specialty workshop or a niche technology discussion group here they can talk about it. That can be as useful as a press release in the early market.  

yegg: And they are willing to do this because they feel they are getting a very good deal?
skmurphy: Yes, they are typically gaining some kind of significant competitive advantage, and they believe they are early in a technology cycle where the startup will continue to invest and develop for several years to come, so they are going to see increasing returns.  

Now, if you sell more copies, those copies may go at higher prices, and as you proliferate in larger firms, things look different.  I think where people get hung up is it's a huge thing to get people's attention, to actually get their data, to have them give you a real evaluation of what's wrong with what you are doing, and to actually solve some kind of production problem using that technology.  

We've been at this now seven years.  I think maybe once after a pilot project, we weren't able to close the deal. I mean where the pilot project was successful.  Obviously sometimes, not everything new works, but for the most part in the B2B side if you can make a pilot project work, you can move things forward.  

yegg: What about the #s are the beginning of the funnel? How many warm leads does one need to hit before finding one that will do a pilot project?
skmurphy: As a rough rule of thumb maybe twenty or thirty. If you've got a theory of who should find value in the application, and you talk to twenty people or thirty people, and no one wants to do anything else, than you've got to change your theory of who is going to get the value, or you've got to change something about what you are doing.  

And, you are obviously making adjustments after every conversation or two based on what they are telling you, adjusting for what's not working and what you've left out.  Our model for a lot of the sales materials and the marketing materials we help people develop is that we are spinning that after every conversation or two.  We don't tend to invest a lot in the beginning in glossy brochures and expensive collateral because it's going to change very rapidly if you act on what you are learning.  

yegg: Suppose you don't know the vertical. Do you set up a system where you have a few leads in a few verticals and then iterate from those conversations?
skmurphy: Yes, we can sometimes fire a spread: we might try two or three different things. For that approach we will develop what we call 35 cards.  It's a tiny brochure that fits on two sides of  a 35 card.  It's got a pain question on one side with two to three bullets that substantiate or talk about your ability to address that problem.  The other side has your 15 second elevator pitch.

We'll do one of these for each of the different segments we are thinking about, or even in a segment there might be two or three different pain points that we are trying to explore this way.  At networking events or in other kinds of conversations, when somebody asks what you are doing, we'll hand them one of those, and then talk about it.  The form factor, which is not much bigger than a business card, is comfortable to hand out even in a standup networking session, or in a technical conference setting.  

It doesn't feel awkward the way even a half sheet or full sheet of paper would. And they'll typically put it in their pocket or their purse and keep it.  

When you've actually got permission and you've got a meeting with a person, you are going to have coffee with them or a meal, then we'll do a one pager--an 8½11 piece of paper--that both guys can draw on.  

But, in the very beginning when you are still in discovery mode, trying to figure out what the pain question is, we'll often use 35 cards. At a conference or event you may talk to many people for a few minutes, and leave them with a 3x5 card: it makes a nice summary of your message and has contact information so that either the person you talked to may call you or they may pass it along to somebody in their firm who may call you back.  

The other thing is a lot of the folks we work with are fairly introverted--a lot of successful engineers didn't go in engineering because they wanted to stand up in a room full  of people and give a speech. The 35 card does a couple of things.  It crystallizes the key points they need to make and communicates them succinctly. It helps get the conversation started, and we focus mainly on coaching founders to ask a couple of questions and try and learn more about the prospect's perspective. 

Sometimes their temptation is to start talking about all the things their product does as if the listener has an hour.  Unfortunately you really only have about  ninety seconds to three minutes: if you haven't said something which has  triggered some interest, you've really got to politely let the conversation end. 

yegg: Back to the first customer. When you close the sale, do you push for a real contract?
skmurphy: Yes, we want a formal written contract. Depending upon the vertical, there is typically a mutual NDA that covers the data that they are giving you plus what you are giving them in terms of documentation or other kinds of materials, training materials, product roadmap,...  There is a real written proposal; they are normally going to have a purchasing terms and condition thing you've got to agree to.  You are going to give them a real software license to sign.  You really execute--and debug--the entire process of making the sale.  

yegg: At what # customer does this start to change?
skmurphy: It's less about a particular number and more about having clear criteria for what constitutes a prospect you are confident you can add value for. For example: we are looking for chemistry labs that have at least two dozen bench chemists that are in an industrial setting, but not at a University.  When you've got clear selection criteria that have been validated by closing paying customers then things start to change: you  move more into a regular marketing and sales mode.  That may take your first three customers, your first six, or more.  

yegg: How do you manage the consulting piece across these first six customers?
skmurphy: We are always trying to run a common roadmap, a common feature roadmap across all the customers.  You don't necessarily share the entire roadmap with each customer, but the team has got the set of possibilities they are using to manage development.  If you see it moving in four different directions, you have a problem.  

We worked with a team a few years ago that had sold to four different firms and they called us in because they couldn't see a pattern and were not sure who to call on next.  Essentially they had sold to a butcher, a baker, a candlestick maker, and a blacksmith.  We interviewed all four and found a commonality, but it wasn't a useful commonality to exploit.  So, we picked one of the four types where we felt they added the most value and then we went after just bakers.  

At that point you can then prepare a frequently asked questions list.  You are starting to be able to calculate the prospect return on investment from the technology much more accurately.  As you particularize who you are going after, you are much better able to calculate the value. And that's when it starts to get easier.  

yegg: Does that mean you are telling no to these first customers at some points?
skmurphy: Yes, you are continuingly negotiating. It's not a contracting assignment where you ask, "What would you like us to do for you? We will do whatever you want."  You have to focus on a productized offering, and the prospect has to understand that and want a product as the result of working with you. This is not meant to be custom development that's work for hire: the prospect understands from the beginning that they will sign a software license and you've got the right to resell the offering to other firms.  

yegg: In that light, what makes a good or bad early customer?
skmurphy: Let me describe a couple of subtle failure modes that we spot early now because we see them across many teams, but that the team can have difficulty spotting.  

One is when the prospect invites you in and asks you a lot of questions about the technology, and about the problem area.  You have a two hour meeting in their office, and you've covered the white board three times with information that you have learned, and they are delighted, and they would love to have you back. And the CEO or CTO or the full team wants to go back. They really had a chance to talk about their technology or what they have learned about the problem. They feel appreciated and it really seems like things are going very well.  

In the de-brief we explore what we learned about the customer's problem and solution constraints. As you analyze the prospect's questions, what they've been doing is they've been mapping the market or mapping the technology, but they really haven't really been turning over any or their cards: pain points, constraints, needs...  

But you have been invited back, to you have a second meeting that also great:  they ask a lot more questions and the team feels really appreciated.  But the de-brief shows us no closer to any real understanding of the customer situation and you realize they have been politely deflecting most of your questions in this regard.

At this point you realize what's going on is you are giving away free consulting to a team or a company that really has no interest in buying what have or will develop. They would like to learn a lot about this emerging technology area, or this problem area, or something like that.  And, that happens ten percent to perhaps twenty percent of the time.  

An entrepreneur or team with a bona fide track record in a particular area has to be on guard about this. The acid test for that is the customer never talks about the particulars of their problem.  Even when you ask them questions about how they have solved the problem, how they are managing the problem now, and what options they are considering to improve their solution. Or if you ask them "What are some constraints we've got to be aware of if we propose a solution to you?"  If they are not able to talk about what's going on and what you need to do, you will never make a sale.  

The second situation that's also a waste of time is when someone claims to be a "change agent." He will tell you that your offering is going to have a huge impact; it's going to transform all of General Motors, for example. Substitute your favorite lighthouse customer. Before you get started doing everything that he is telling you to do, you need to ask him, "Have you ever brought other technology into your company?"  More often than not unfortunately he will say "Well no, but you know I've only been here six months, and this is what's going to let me make a big difference here." Or perhaps he has been there a decade and has never brought anything new in but he is bored. 

So then you have to ask him, "Please explain to me what the process is going to look like for you to be able to get this signed off, what do we have to do to help you make that happen?"  He may say he is not really sure. If you are working with a technologist then it's completely legitimate that they don't know.  But, that means you are not working with somebody that's ultimately going to be able to buy the product.  They have to be willing to introduce you to somebody, whether it's their manager or somebody in other department that says yes, I understand the impact on the business, and if Joe says technically it works, we'll figure out how we ultimately get this deployed.  

So, the two typical problems are, you end up giving away free consulting or you talk to somebody that in their own mind is this change agent, but they have no idea how to make it happen.  And, most of the time a real change agent would be very candid with you about what they can and can't do and what you've got to do to get the business.  And it's not so much a negotiation as explaining the constraints the organization has.  

yegg: What about the case where the prospect wants to buy a product but they don't want to get into developing a product with your or doing testimonials?
skmurphy: So, you hang onto those guys and you come back to them when you are able to demonstrate value very rapidly, and in a demo or a brief benchmark say here it is, here is the price, would you like it?  That's the start of your "regular sales process." Those people tend to be more pragmatic or even late adopters.  You never want to turn them off, you just want to say, let me get back to you, it seems like we are not quite solid enough, or stable enough yet. Talk with them enough to understand their needs and constraints on buying and then nurture them until you have some other customers.

Blank has got this earlyvangelist formulation where they have the problem, they know they have the problem, they've tried to solve the problem, they are unhappy with their solution to the problem and they can get budget.  That's a very good checklist.  

yegg: What about the customer who seems good--you've signed the agreement and everything, but they start demanding things you aren't really going to do? They want the consulting piece to be much heavier.
skmurphy: There are a couple of ways to solve that.  Big customers, early customers are always going to be demanding. You need to appreciate that  your understanding of the real requirements may be flawed or incomplete.  What sounds initially like a unique requirement at XYZco may be an actual market requirement, or at least a market requirement for a segment of the customers you are targeting.  

One test you can use is whether you can sell it to three other customers. Ask them "Are you asking me for something that's unique to your organization? Who else will have this need?" If you've got a good working relationship with them you can you can phrase it a little more artfully than I just did, but essentially that's the test.  

If the answer is, it's just for us, then you carve out a consulting rate on that and you do a special.  And, for your first couple of customers, for extra money, you may do some special work. Depending upon the market specials may always be a component of your business for ten to twenty percent of revenue. 

You don't want it to get above a third of your revenue because you will drift into becoming a specialty application developer.  If more a third of your revenue is coming from customer consulting, you are probably starting to have a problem.  

For your first one or two customers that may look more like 50% consulting. But you've really got to be driving to a product that a direct sales or channel sales guy can sell. 

yegg: At what point do you advise bringing in a sales person?
skmurphy: If you've made between a half a dozen to two dozen sales and  you've got a sales presentation that you know works.  If you've got a way to target prospects: you understand if they answer yes, no, or a number to these six questions, you know half the time they are likely to buy.  You don't necessarily win half the time, but you've got a reasonable presumption you can create value for them.  

Then you know you can fire a sales guy if it's not working because the problem is him or her, it's not your product or process.  What often happens is that the founders add a sales guy, before they can manage the sales process. They think somehow that guy is going to take his rolodex and solve all their problems.  They have to get--at least one member of the founding team has to get--good at understanding the customer negotiation process, and what's going on there, so that if the salesperson is bullshitting them, they can call them on it.  

And this is nothing against sales people: they have a critical role to fulfill. Most successful sales people won't come into a situation that's too early because they can't make that much money.  If the product doesn't work well, if you can't tell them "Go talk to people like this, look in your rolodex and find all the left-handed tuba players that used to live in Cleveland", then they can't make money, and they don't like that.  

yegg: For your typical engineering founding team, do you advise one of them to develop these skills, or go find another co-founder, or what?
skmurphy: Most of the time at least one of the folks is more outgoing and is willing to get out there and to take part in the process.  If everyone on the team believes that they can stay in the BatCave and use Google AdWords to make their money then they better have a product that matches that profile. It  is typically not the products at the price points we are talking about, where we're doing annual license costs of anywhere from a thousand to fifty thousand dollars a seat, where deal sizes are in the twenty-five thousand to quarter million dollar range.

There are other markets where the founders don't have to as outgoing.  You had Patrick McKenzie on; his whole business model is B2C: he and I are on different planets.  That doesn't mean that I am right and he is wrong or he is right and I am wrong.  We are just going after very different kinds of markets: I want to be very clear that I am talking about business-to-business software sales. 

yegg: Is there a typical person on the side, e.g. title, that you are dealing with?
skmurphy: It's typically a midlevel person who has got a real problem.  It may be a first-level manager, may be a senior contributor, it could be someone in the finance, it very much depends on the product.  When we help sell medical workflow products we call on very different people than when we help  sell legal services automation or computational chemistry tools.

But ordinarily, it's somebody who is one level or two levels up in the organization; they've got enough perspective on the problem and on the organization to understand what's going to be involved in bringing change to the organization.  As we work with them they may take us up the hierarchy to sell more senior folks.  

We don't tend to start at the top unless we are calling on a very small business in which case you've got to call on the CEO or one of the key execs because no one else can make any decisions.  

yegg: Suppose you hire a salesperson. How do you approach that process--do you hire just one to start?
skmurphy: Yes, start with just one. And hire someone who wants to sell, not a sales manager who wants to hire sales people and manage them. 

Sales guys are very expensive, but you have to realize that if they can't make money then they can't make money for you. They face a learning curve process on your product:  you've got to develop real sales training material, which is different from the customer material that you've already got developed. 

You've got to be able to teach them how to sell your product.  Now, you don't have to teach them how to sell--how to navigate a customer hierarchy or coordinate a sales process or orchestrate a sale.  You are not giving them sales training in the abstract, but you need to give them training on how to sell your product.  

For example: what does it mean when the customer asks this question; what's our standard pricing and discount structure, how can you disqualify a weak prospect so that you focus on strong prospects.  

yegg: Do you typically advise one of the engineers to become a pre-sales engineer in this stage?
skmurphy: Ordinarily, they are going to have to.  Now, and we'll also go on sales calls, so myself and my partners will also go on sales calls in Silicon Valley.  We can help with this. It's less stressful to be sales engineer than the sales person. But both people are involved in the sale, they each have to be able to read the situation and improvise to assist the other.  

The biggest challenge that a CTO or a technical person faces in that setting is because they've developed their product with a particular perspective, they will sometimes tell the customer, you know you are using my product wrong, stop doing that.  And, sometimes they'll almost say those exact words.  And, the reality is many times the customers "misuse of the product" is actually where a bigger market is.  Maybe they are only using three of the seventeen features, but they are creating more value and you've got to be alert to that possibility.  

yegg: The salesperson that you hire first--presumably you want them to have domain expertise. Are there other parameters that you are looking for as well?
skmurphy: In the markets we operate in we like them to understand what's involved in an orchestrated sale.  The early sales are still going to remain complex; it's not going to become a transaction for a while.  So, you need somebody who has done that.  It may be someone that has ambitions to be a sales manager, or may have been a sales manager who was willing to take a step back for a year to be able to then help you build a team.  

One other mistake we see folks make, not so much the bootstrappers because they just don't have the money, but we'll see teams with venture funding that'll go hire, you know a VP of sales, two sales managers, four sales people.  They don't have training materials; they haven't figured out how to manage a process; and you create this train wreck and you burn through a lot of cash.  

yegg: So in either case (bootstrap or venture), you're saying you should definitely start with one person, get the process down, and then bring in more people.
skmurphy: Yes, and referring to your earlier question about titles, there are going to be a few titles you are going to be calling on.  And, you would like to know that this salesperson has called on those kinds of people, and perhaps has some in their rolodex. 

yegg: Let's back up to the founding team. What typically will the founding team consist of to maximize success probability in this arena?
skmurphy: Two to five people, engineers or scientists that have got relevant work experience in the domain or the industry that they are trying to sell to. It's okay if only one of the two or two of the five, or one of the five has got domain knowledge because sometimes bringing people from different fields is useful. 

Outsiders can create strategic surprise because they bring things that work from different fields to this new field.  It's not a requirement that everybody has worked in the field you are trying to sell into, but if no one on the team has worked in the target industry or field then you may be missing so much basic context that you just can't get there.  

yegg: Do you see that in real life? Entrepreneurs seem to do that a lot--approach an entire new industry looking for problems and hoping they can solve them within it.
skmurphy: I think that's good for an "Act Two" where you've got a proven solution in a primary industry.  I think as a startup trying to make your early sales in a new industry, I think the credibility problem can be very, very challenging.  The only way we've seen that work is where you are able to very quickly ask the right questions and demonstrate some capability that people can see the benefit for themselves.  

There is also a shibboleth effect: you may not be aware of how people in a domain use certain words in a way that's very different form their everyday use or understand jargon or terminology that only has meaning in the industry. You may mispronounce words, use a term the wrong way, or misunderstand someone. This will put a prospect very much on edge and suggest that you don't know what you are talking about.  Sometimes there are dialects, especially in a very early market, and people from different companies or traditions will use different words to mean the same thing. So you also have to appreciate that and not try and correct the prospect. 

Ordinarily we suggest people start with something they know pretty well, and then branch out.  Now, the other way to do that is to bring somebody else onto the team that knows the domain you are going after really well.  So, just to be clear, I am not saying that the entire team has to have deep domain expertise, but if no one on the team does, it can be very hard.  

yegg: What else am I missing about this process?
skmurphy: I am giving you an astigmatic perspective, distorted by our focus on bootstrappers in B2B who typically have a decade or more of experience.  We do work with younger teams but even there at least one has a couple of years of domain experience. Other markets work very differently from what I can observe. We are fans of Steve Blank's "Four Steps to the Epiphany" which is fundamentally a B2B model.  

I think one thing that we tell founders that seems to be the most baffling for them is "We are not going on a sales call, we are going to have a conversation; we are going to be genuinely interested in the person's problems learn more them."  The teams that tend to be more successful are motivated to solve the problem as much as develop a particular tool they've got today to solve it. 

For example, whatever you start out with in January of 2010, by the middle of 2011 your solution is going to evolve considerably.  Too often there is "better mousetrap" thinking: the entrepreneur believes he has this fist-sized chunk of kryptonite in his trunk, and people with Geiger counters will find him in a parking lot and buy a piece. It doesn't work that way.  

I think the other challenge that new entrepreneurs have trouble with is that it's as much about self-improvement and developing new skills as it is taking what you've developed and finding a market for it.  I mean it's a very difficult journey and you have to change if you want to continue along it.

yegg: How long does the process of getting the first stage of customers typically take?
skmurphy: Let's assume for the moment that we are doing some amount of rehearsal before the conversation; that we are making small modifications to the one-page or to the material, later on it becomes maybe an eight-slide deck, ten-slide deck at most. You can do maybe two of those a day if you are working really hard.  

Realistically, it takes a while to get on people's calendars.  You are really trying to talk to people that are legitimate proxies for who is going to ultimately buy it from you. So, that may take two to three months to have those conversations, learn from them, and be where you are.  And, there are sometimes ways to accelerate that by going to what we call target-rich environments, by going to the right conferences, by going to the right places where you can talk to lot of people.  

But in the very beginning telling the same story to ten people is a lot of times less efficient in the long an than if you took the time reflect on the interaction, and make small improvements. You lose that ability to learn and adjust if you try and go too fast.  

yegg: Would it be then fair to say that the first year is about securing those first few customers and getting it working on-site?
skmurphy: That's a reasonable timeframe: six to eighteen months to have a couple of customers really in production and some testimonials. Part of the reason why we so tolerant of consulting is that it allows you to sell a mix of things and your product gradually takes over the mix.  

yegg: Then roughly, by the end of the second year, you would expect to have defined the sales cycle quite a bit, and potentially hiring a sales person at that point?
skmurphy: Yes, it can vary but somewhere between nine months and eighteen months, if you've been able to close business,  you can definitely start to interview people and to look to bring someone on.  

There are also ways to leverage partners as intermediary.  The next step up sometimes can be your product actually works well with other products.  You've got customers in common; they find it advantageous to bring you in because you solve problems that then create opportunities for them.  So, there is also the possibility of other kinds of solution partners or channel partners.  

yegg: In the channel partner case, are there any particular gotchas or ways that you like to frame those agreements?
skmurphy: So a pure sales channel is actually harder to bring on then individual salesperson.  You should bring on an individual salesperson first because you've got the sales training problem magnified by a factor of ten or a hundred.  Whenever you are negotiating those agreements, figure out how you are going to get out of them; figure out what constitutes good results and how do you terminate if it's not working. Spend some time up front on how both sides are going to keep score on the agreement.  Be careful of exclusivity: if they want exclusivity, there is going to be some kind of guaranteed payment or guaranteed revenue stream.  

So ordinarily we would see some kind of solution partner or somebody else that can help, then of a single sales person, from there you may add a second sales person, you may look for channel partners.  A lot of companies end up in the four million dollar to ten million dollar level, and if they've got a recurring revenue model then one or two sales guys can keep it going.  

yegg: On the bootstrapper side, is there a particular way you look at this as the right time to raise outside money or not?
skmurphy: When you go to raise money you need to be able to explain to investors how you are going to pay them back.  If you really are held back by an inability to take advantage of the opportunities you've uncovered, if you've got a real recipe for making your business work and how to scale sales up, then that's a good time to raise money.  

But, if you are running out of money and you can't meet payroll, and you can't quite figure out who you'll sell to next, taking six months out trying to raise money will probably kill you.  I mean a lot of people that are running out of money think that finding investment is the answer, and that's almost never appealing to an investor.  

yegg: Putting that in the context of the cycle we've been talking about, it seems you'd be raising money when you're a little further a long, i.e. after the first few customers and you're ready to bring on the first sales person?
skmurphy: Sure, it's legitimate to seek investment to be able to pay the sales guy if you have a proven sales process one of the founders can manage. A brief detour on commission only sales people: it seems like it's a way to save money, but in general you don't get their full attention and it doesn't do you a lot of good.  

yegg: And how much would you price an initial sales person?
skmurphy: As a base, again Silicon Valley, recessionary times, I mean might be 5k a month, might be 4k a month--depends on the person sometimes, a little less.  There might be a non-recoverable draw of that much again for the first three months.  The real challenge is you've got to be willing to fire them within sixty days, ninety days if it's not working out.  And, you've got to have a plan for how you are going to evaluate how good they are, and go on to the next person if this guy is not working.  

yegg: And that would presumably be you've set aside some leads, you send them on those calls, and see what happens?
skmurphy: It's still a team effort, you go with them; you keep a close eye on them.  It's very different from an engineering hire where you can look at the guy's code and make an assessment of what's going on. Sales is primarily improv.  You can look at the proposals; I mean there is some amount that's written.  But, one of the founders has to be going on the sales calls.  

yegg: But ultimately, even if you like what they are doing on the calls, if the business isn't closing, it is probably time to look for someone new?
skmurphy: Well, and the presumption there is that prior to bringing them on, you had been reliably closing business. Now, reliably closing business may mean it still takes you six months from the time you first talk to the person to close the deal.  The sale cycles can be long, and I am not saying you've got to fire somebody in ninety days if they haven't closed a sale, but you should be able to see bona fide progression markers pretty quickly.  You are moving from initial calls to customer supplied data to evaluations to final proposals. 

I am not a fan of probability of close metrics. I like to see sales milestones that are tangible: the prospect has to have done something in the last two weeks to indicate interest or progress in making a decision. 

yegg: What else?
skmurphy:  I want to thank you for the opportunity to do this interview.  It's helped me to clarify my thinking in some areas. I appreciated your questions; they have really made me think. Thanks again. 

p!=t, i.e. plausible != true

Conspiracy theory cat.jpg
Just because something is plausible, doesn't make it true. 

  • News. X happened and reporting says it happened because of Y & Z with seemingly no data to back it up.

  • Finance. Some market does X because of some other event Y, or not.

  • Science. Study shows X but headline and reporting around it extrapolate to Y & Z.

  • Conspiracies. No example needed :)

  • Startups. Perhaps most close to home...awesome product idea X is going to make a killing in niche Y because of Z, but no one has talked to any customers yet. 
etc. etc. 

tl;dr comments really bother me when they don't have a useful summary attached to them. I don't think I'd have a problem seeing p!=t comments, however. 

Long articles are, well, long. But p!=t articles are dangerous, and even if they have kernels of truth in them, I think they do more harm than good.

I have nothing particularly against probabilistic decision trees and scenario planning, but it should be highlighted where you go from fact to speculation, and to what degree that speculation is probable or not.


False dichotomies in convertible note vs equity seed rounds

Using convertible notes vs priced equity in seed rounds is a hot topic right now. For background, see recent posts by Seth LevineChris Dixon, Fred Wilson, Paul Graham, Ben Yoskovitz, Bill BurnhamMark Suster, William Carleton and Patrick McKenzie as well as older posts by Venture Hacks, Yokum Taku, Brad Feld, and Josh Kopelman. If I missed any good ones, please let me know and I'll add them to this list.

In an attempt to usefully add to this discussion (instead of repeating things from the above), I want to highlight what I think are false dichotomies between convertible notes and priced equity rounds. FWIW, I've done four deals so far (including both types), and between each of those reviewed a ton more.

Control vs no control

You can do priced rounds with no control (just economic rights) and converts with some control (e.g. converts into preferred shares w/ control provisions on maturity, stipulations about early acquisitions and next financings, etc.).

High-res vs non-high-res

High-res in this context means giving different investors different prices based on their risk taken and value added to the company. You can do priced rounds at different share prices or converts with different valuation caps. In practice I currently see more high-res in priced rounds via advisor shares. In that case, you reward the more valuable investors via options to be dispensed over time if they continue to add value.

Cheap vs expensive

You can do both on the cheap (or expensively). All my deals so far have had literally zero legal expense on my side. Use the standard docs and you're set.

Simple vs complex

You can do both in a simple or complicated way. Again, use the standard docs and it is simple. Personally, I find converts to be more complicated currently because there are no good standard docs yet and there are a lot of edge cases, e.g. what happens on acquisition, no Series A, a non-qualified financing, on default, etc. So I therefore find myself having to carefully review any convert wheras I'm ready right now to sign the series seed docs.

Better vs worse (for founders or investors)

There are valid reasons on both the entrepreneur side and the investor side to prefer doing a particular seed round with either a convertible note or priced equity (more on that below).

Quick vs long

I haven't perceived any difference in time to close. 

Fixed amount vs variable amount

Priced round term-sheets generally have a range of money you're willing to raise. You can easily make the docs such that that high number is high and leave time to add more investors ad-hoc.

Valuation vs no valuation

Converts (mostly) all have caps now, which is effectively a valuation. At least that's how I view it.

The debate of convert vs equity has seemingly turned binary, which is hardly ever a great approach to anything. Instead, it should depend on your particulars. One particular that I think matters more than the others is what probability the company is planning to do another round of financing and in what time frame. 

For example, if you're planning on definitely raising a series A within 6 months and the seed round is really a bridge loan to de-risk an assumption and get a higher valuation, then that is a decent argument for using a convert. On the other hand, if you have a decent probability (say 50%) of never needing another round of funding, or you think this seed round will last you 18+ months, those are decent arguments for doing a priced round.

Thank you to Roy Rodenstein & Jeff Miller for reviewing this post.