Tell me what you are going to do tomorrow!

Entrepreneurship has never been easier: Buy Steve Blank’s book on Customer Development; read all of Eric Ries’ posts; reduce your fundraising presentation to the Business Model Canvas; wear a hoodie and pivot like a lean ninja. Success is guaranteed.


Creating value from nothing is just as hard as it has always been. You wouldn’t get that from many of the investment pitches that I see though. Instead, I hear about “viral” marketing plans without budget (but massive predicted revenue), “lean” business models without unfair advantages (but massive predicted profit) and “Ramen” operating plans without paid employees (but cheerful ever increasing productivity). Don’t fall into this trap. Focus on real actions, not fluffy jargon. Instead of “viral marketing”, tell me that you are going to put short in-progress videos of our product on YouTube every week. That sounds less sexy but at least it gives investors something factual to evaluate and you an actual task.

Similarly, don’t spend your day doodling on the Business Model Canvas in a quest for the ultimate business strategy. Strategy is important, but it won’t come from using a(ny) tool. Recognize that these tools will at best help you to organise your thoughts. Don’t mistake them for answers to the unique questions in your venture. For that you need to go out into the world and get things done.

Get out there even if it hurts!

Over a decade of entrepreneurial activity in “hard” technologies has left me with a deep repertoire of demo horror stories. I have had boxes blow up, catch fire, over-heat and over-cool, human failure, device failure and some that felt like divine failure. You name it, I’ve suffered it. Maybe that just means that I am a bad technologist, but I like to think that it is the result of a two personal mantras instead:

Perfect is the enemy of “Good Enough”

Get feedback early and often

I have always believed that it is better to get customer or strategic feedback on a new technology *before* you have turned it into a complete product (“complete” has no well-defined meaning in a disruptive new technology in the first place). Early feedback allows you to optimize your development strategy and might even bring in some initial revenue. Steve Blank and Eric Ries have turned this approach into a science of sorts, so I won’t belabour the benefits of customer development. It sounds great and works great. Except that you are leaving the building with incomplete crap.

The customer development crowd talks about the latter aspect a lot less – partly because of their focus on the web/software world. You can mock up a website and demo a big portion of the concept. Not so with hardware! There was no viable “photoshop mock up” of LED TV when we first pitched the concept. For starters, we were promising unheard of image contrast. Showing that on any current projector, TV or laptop was akin to advertising colour TV with black-and-white monitors. We ended up building “simulators” with a modified DLP projector behind an LCD panel.. Optically that was the equivalent of future LED TV so we got a “mock-up” of the visual quality achievable with this new device class. Except that pitching the “next flat panel TV” to potential customers/partners/investors with a 6′ deep demo requires quite a dance (we called the demo the “coffin” for a reason).

A lot of extremely talented inventors lack the confidence to show incomplete demos. I see this a lot and my usual response isn’t a customer development lecture but a story from my own demo days. These stories, while often scary, hopefully provide the inventors with a bit more confidence: I did this, many things went horribly wrong, but the end result was still good.

As I am learning from blogging masters like Mark Suster, blogs are a place to store these kinds of “often repeated stories”. Thus a new category of “Demo Wars” is born. I will start with a quick one and add more over time. Maybe these inspire some of you hardware folks to go out into the demo world.

Keeping your Cool

Shortly after the “Coffin”, we puzzled together our first true LED prototype. We used LEDs used in flashlights for lack of other options. Cutting leads and soldering 2000 LEDs left me with a permanent aversion to electronics work (and likely some brain damage from the fumes…). Absent dedicated LED control chips, we had the entire system run on analog designs. That design, and I am sure several implementation flaws, made the display as temperamental as my two year old. Specifically, the LEDs would start to flicker and buzz like crazy at any temperature below 40C or above 45C. For a display sucking over a kilowatt of power and using only “hope this work” cooling, that temperature band is absurdly small and the failure mode was truly spectacular.

That’s what we had. We also had an opportunity to showcase our invention at the largest technical conference on display. Of course we went!

Our talk at the conference generated a lot of interest so we had a crowd waiting at our demo booth. This is when it got tricky. I would put my jacket over the unit while people were gathering in anticipation. Holding the jacket allowed me to feel the rising temperature of the system (jacket blocking airflow). Once it felt like 40C or higher, the jacket would come off and everybody went “wow”. This would last for about a minute and then the display would push past the upper limit of its thermal envelop. For that scenario I had a bottle of spray coolant in my other hand behind the display. A gentle spray and the unit would drift down a few degrees again. Overshoot the coolant and the jacket came back on. And so forth. For hours…

I was a nervous wreck afterwards but learned my first really important lesson: Nobody commented about the stuff I did behind the curtain (or behind the jacket as it were). I think most people didn’t even realise what was happening and others just didn’t care. I even got the odd questions about “the next exclusive showing” (i.e. the jacket being removed). Instead, we have hundreds of display industry leaders gather in front of the display over time. The device had a fairly narrow viewing angle of about 30 degrees and I still remember seeing a tight crowd of people wedged tightly into the narrow cone, standing up to 10 people deep all the way into the booth of the booth across the hallway.

Go out and get feedback – no matter how awkward it might feel!


Customer Engagement & Metrics

As promised, here is the first detailed follow-up to the introductory post for the topic of university spin-offs versus conventional start-ups.

One of the most common recommendations for entrepreneurs is to focus on customer engagement early in their new venture. In fact, a whole set of new online nomenclature is devoted to this concept: Have an idea, put it together quickly to get customer feedback and then iterate based on good metrics. That sounds great, and is great advice when you are building a Web2.0 start-up. The problem is that most university spin-offs simply cannot make this work in the way described. First, core technology often requires considerable investment before there is anything to show, much less a product to put in front of customers. Mashing up a new way to share photos with your online friends can happen over a weekend. Developing a quantum computer, new pharmaceutical or microchip can’t.

The second problem is that the customers for university inventions are hard to approach and even harder to assess. Not only do you have to physically move around prototypes, your “customers” will likely be executives of large corporations. Even if you can get to them, there is an inherent bias in the responses that you will get. A Web2.0 user might like or dislike your product but won’t ever be a competitor to you. When you are commercialising core technology, you will constantly be faced with customers who are at the same time competitors. Worse, a lot of university innovation is disruptive, so the bias can be extremely strong.

I have experienced the impact of this problem more times than I can count. I remember introducing our idea of putting controlled LEDs behind an LCD to create a new type of high contrast display. My very first audience question after the talk, in front of hundreds of industry veterans, came from a senior executive of a large consumer electronics company: “This concept is not viable for the consumer market at all for several reasons”. He then went on to enumerate a number of reasons that were challenging to validate for us at the time but sounded intuitively believable (e.g. internal cost of manufacturing pick & place operations, etc.). Less than two years later, that same company introduced the first commercial LED-based LCD display – a product that beyond any doubt was under active development during the time of my talk. Fortunately we applied the right filter to this feedback (and many other comments like it). Today, that manufacturer and its competitors are shipping tens of millions of LED TV. We stuck to our guns and had a very successful exit but it could have easily gone the other way if we had listened carelessly to our “customer feedback” during the early days.

Of course this doesn’t mean that you shouldn’t try to validate your technology and business model. Never stay in the cave! But it does mean that you have to find a different way to go about it. There are a few options to replace traditional customer engagement metrics: Advisors and Trend Indicators


In my experience the best option is to build up a network of strong advisors. Don’t look for believers who are already championing your new idea (or easy converts). Look for people who are domain experts in the field that you are entering and who have a perspective on the challenges in that space. If you can convince those people that your concept is viable then you have just gotten a major validation for your path. Throughout my career I have always maintained a strong advisory board and the impact of these groups has been dramatic. Ranging from deep technical experts to retired senior executives of large companies, they have helped to shape the path for the various concepts that I have worked on (as well as being a source of technical inspiration and innovation).

Trend Indicators

Another option is to use academic conferences as trend indicators. This is a lot more complicated but worked very well for me in the past. Try to identify two to four major fields that impact your core business. In the above mentioned case of our LED backlit LCD technology, those fields were LCD (more specifically the brightness and contrast of LCD), LEDs (their cost and efficiency) and Green Technology (the push for lower energy consumption). As our venture progresses we kept very close watch on the statistical behaviour of the industry. This can be easy or hard. For example the cost of LEDs was well known and even had a good forecasting method (Haitz Law – LED prices fall by 10x every decade).

Other aspects where a bit more challenging to quantify. Contrast of display is a major battleground of the spec warriors today but in 2001 it was still far into the future (“thin” was the big deal then). So I looked at some of the major industry conferences and specifically gathered the contrast and brightness performance of top displays each year. Taking a blend of award winning products as well as prototypes described in award winning papers, you could create a past history and future roadmap for display contrast/brightness. I dug up my own data for this post to serve as an example. The blue dots are the blended average of top displays of each year that I actively collected this information. Somewhere around a contrast of 1000:1 to 3000:1 you start to hit significant physical limits for LCDs and a secondary source of contrast enhancement is needed (i.e. our idea of controlling LED in the backlight). We shifted our business strategy from high end professional products to consumer licensing in 2005, largely on the basis of trend lines like this one (a huge strategic discussion and risk!). I added the red points after the fact (we sold our company to Dolby Labs in 2007 – they could read this graph too).

In summary, validating your business model is always a good idea but doing so requires often a bit more finesse when you are working on core technology. Fortunately, one of the benefits of being a university entrepreneur in a spin-off is that you have often a lot better access to experts and conference data in your field (compared to a start-up founder). Use these to your advantage and they can fill the gap in the customer feedback loop.