Category Archives: Research

Continuous Visible Customer Validation: Using Do, Doing Done to Customer Development

There are multiple methods to keep track of your projects and priorities. Here is one technique I used when trying to keep track of customer validation. This approach works best when teams are in one single location since it is visually appealing and easy to update.

It is called the Kanban method of Do, Doing and Done. A visual dashboard, which I used colored post it notes for, comprises of 3 buckets of work:

1. Do – has all the items that need to be still done.

2. Doing – what you are currently working on

3. Done – what’s finished

Typically we update and refresh the items each week – sometimes daily if there are enough changes, but a week is ideal.

Do Doing Done Kanban Method
Do Doing Done Kanban Method

Some folks have mentioned using Trello or Kanbanize. I have not, so I cant recommend them.

Once you understand how to segment your startups customers and you understand the 3 most important steps to segmenting your customers, most people start to put a framework for validating customer segments.

As I mentioned previously, the elapsed time for these 5 steps, in my experience lasts from a 4 weeks to 3 months on average.

The most important item to remember is that the method works best for discrete, defined tasks that take a short period of time. If your item takes many weeks or months, you have to break it down into simpler steps. Customer validation or pricing strategy cant be a step. Pricing strategy has to broken up into price tiers, price testing, pricing validation, pricing research etc.

When validating customer problems, you are trying to understand the following questions:

  • Is this a real problem? Is is a big enough problem for them to look for a solution?
  • What will it take for them to adopt a solution? Adopt my solution?
  • How much will they be willing to pay to adopt?

I have used post-it notes as a great way to segment the steps in the process and use colors to validate different items I need to:

In the customer validation I put forth a process comprised of 5 steps:

1. Secondary research

2. Primary research with insiders

3. Proxy market sizing

4. Online validation

5. Customer interviews

The best way to use color is to put these various “sources” into different colors. So, influencers may be blue, proxy sizing sources may be yellow, etc.

When you get to the customer interview step, you are likely finished with the previous steps, so you can color code segments of your customer with different colors.

So “Segment A” will be yellow and “Segment B” will be green and so on.

The reason for colors is then you can put them all into a board at the end and find a way to look for patterns that correlate.

The big advantage of the visible productivity map is that everyone is motivated to make changes to the charts so you can see progress daily. Try it and let me know if it works for you.

What has changed for developers in the last 20 years

I asked this question on Hacker News last week to understand the shifts in software development over the last 20 years. From 1995 to 2015, there has been a dramatic change in the developer ecosystem. I thought I’d summarize all the changes and try to make sense of the trends. In this post I am only going to focus on the identification of the trends, as opposed to the analysis. I would love your thoughts on trends I may have missed.

1. The rise of open source options: In 1995, there were about 5 open source languages for the web including Perl. Now there are over 100 languages including Ruby, PhP and Javascript.

2. Plethora of libraries and frameworks: From < 10 libraries and frameworks to over 200 (Bootstrap, Javascript frameworks, etc.) The only libraries available in 1995 were those for Javascript. Today, there are over 100 libraries and frameworks for Php alone.

3. From waterfall approach to development to Agile: Most early software development was based on Requirements -> Design -> Architecture -> Development -> Testing -> Release. Now with agile methodologies being followed by many development teams, we are seeing a rise of faster release and in many cases daily releases.

4. Client-server application development to Web apps to Mobile apps: The overall changes are from PC (dekstop / laptop) client software to web applications and now to mobile applications. We have gone from native clients to browser based apps back to native mobile apps all over again.

5. Phenomenal rise of consumer apps, thanks to mobile : Personal finance (Intuit), to 1+ Million consumer apps thanks to mobile. PC’s were largely (90%) used for “work” with few consumers having home PC’s. The home PC’s rose thanks to the web, but now everyone has a mobile phone. Which has led to a phenomenal increase in # of consumer apps, not just business or productivity apps.

6. Increased availability of application level API’s: From providers such as Facebook, Twitter, and others on programmable web. The abstraction of core API’s from just Operating system SDK’s to application level API’s has made the move for apps to be built on the next level of the application stack.

7. Ease of looking up coding examples, tutorials and sample code: Thanks to Stack Overflow and Github, there are many more samples, code snippets and examples that developers can use to be more productive quicker.

8. Rise of coding / hacking schools: From no programming skills to employed developer in less than 6 months. Most developers, 20 years ago, needed to have an education in Computer science, before they could code. With the rise of frameworks and libraries, along with higher level languages, there has been a significant rise in number of coding schools and bootcamps to get anyone with any degree be a developer in less than 6 months.

9. Increase in the number of indie developer (solo): With the rise of consumer mobile apps and mobile games, there has been a significant rise in # of solo developers who are able to make a living based on building applications for niche audiences.

10. The change in market share of complied versus interpreted languages: 20 years ago, most programs and applications were compiled (C, C++) and the share of interpreted languages was small. Now, with Javascript Ruby and Php taking the forefront, most applications are interpreted not compiled. The only exception is mobile apps – which are still compiled.

11. The rise of DevOps: Developers are now being asked to not just build and architect, but also release and push their apps to production. Roles that were previously performed by specialized system administrators and release engineers are now performed by the software developers themselves.

12. The fall of software testing: More developers are also being asked to test their own code and software applications, instead of handing off the testing to a separate team.

13. The changes in app distribution – App Stores: Discovering, installing and using apps is a much more smoother and easier process now than before thanks to App stores.

14. The availability of Cloud infrastructure for app development: The biggest change for developers over the last 10 years has been the rise of AWS and other cloud services, which allow developers to provision, build and deploy instances and machines much faster than 20 years ago.

I’d love your feedback on the relative ranking of these trends and if I have missed any trends I’d love your feedback on those as well.

The “Goldilocks” overview presentation for #startups – not too technical, not too fluffy

Many #developer founders struggle with their pitch to anyone but their customers. Too technical and they end up losing 90% of their audience, like investors or potential employees not in the engineering team. Too high-level and everyone thinks they are hand waving.

The problem is fairly acute in B2B companies overall – if your product is aimed at a very technical audience – for example finance managers, statisticians, or climatologists, then you will end up getting “into the details”, in your overview pitch.

The right level of presentation is very hard to get right. It almost seems likes a “Goldilocks presentation” – not too technical, which most people wont get and neither too fluffy – which many dismiss as “does not get the problem right”.

The simple answer is to keep it on the right side of technical. From my experience it is better to be specific and articulate than come off as condescending or “hand wavy”.

The good thing is that this also will ensure that if some folks in the audience dont get it, they are probably not the right target for you.

So, the question is what is the right level of technical? The answer wont be easy, but the best thing to do is to A/B test your positioning with the soft audiences first.

The most important part to remember is that it is not only investors who are the audience you are initially trying to get on board. 

Sometimes senior executives in your potential customer base have a problem relating to very technical presentations, as well.

If your customers dont get your pitch – again either because it is too technical or too fluffy, then I’d recommend you revisit the lucidity of your presentation.

Let me give a specific example of one startup we are helping now. They target a very new and a developer audience. Most of what they end up doing is “Educating” their audience.

When they talk to potential customers at the right level in the organization, the bells toll, but in many cases when they describe their problem statement to folks higher in the org of their target customer base, things get difficult.

Here is what I recommend:

Most people understand the BEFORE and AFTER story the best for representing technical products.

If you have to explain a trend you might want to articulate that quickly, but I’d focus a lot on sharing what the “CURRENT” problem is – which is the BEFORE situation.

For example. The pitch they were using was to show code screen shots of deployment tools and how their product was much better. That went well with some developers, but they were unable to sell that to the managers who needed to understand how it will help developers.

Most managers, when they did not understand it clearly enough, dismissed the tool as “nice to have”.

Here is a better “framing” of the problem in my mind.

1) Your developers need to understand agile methodology since they are being asked to ship products quicker and in incremental fashion instead of once every 6 months.

2) Developers like the agile methodology but your systems are built for the waterfall approach

3) If you use the tools like abc and def which were built for the waterfall methodology, the compromises they will show up in more outages, more defects and slower release cycles.

This helps put a context to the person listening to the pitch even if they are not using the tool daily.

You will still have to tailor your “standard” pitch so it appeals to the audience, but this is at the “right level”. Again, you want to keep testing, until you can get head nods quickly, within the first 1-3 minutes.

That’s when you know you have the pitch “Just right”.

Apple Watch is going to hurt Twitter the most. Law of unintended consequences

I have been reading the multiple blog posts on the Monday “Spring event” for the Apple watch.

Having worn the Microsoft band for a few months now, I think I now know what I need from a wearable. Note I did not say “watch”. I gave up wearing watches many years ago and switched to a phone for time. I really don’t have a need for a watch and so don’t many others, but they will still buy the Apple “watch”.

I used the fitbit for activity tracking, so I was not actively looking for a fitness tracker before I got the Microsoft band. Being an active user, I think that I want most is “Smart Notifications” from a wearable. That it will track some fitness is an added bonus.

With the very small form factor, it is absolutely important that the right amount of “relevant information” comes to the wearable.

If you just take an email and strip out a few things and send it to the wearable, that wont help.

What you really need is a summary of the relevant portion of the email and the ability to dismiss, delete or provide contextual reply – the relevant actions may differ on the notification itself, but the action should result in not having to pick up the phone for quick responses, which your watch can handle.

Lets look at email notifications first and email call to actions.

I am really surprised the the Microsoft band has no delete or archive actions on the emails received. Which is pretty awful actually. I am pretty sure 60% of all emails that I receive are to be deleted after reading immediately or archived. Of the remainder, I could guess that 50% of them would be able to get a simple answer – Thanks, OK, Sounds good, Approved, etc. I am surprised that does not exist on the Microsoft Band.

The notifications on the Band are not “smart”, which I suspect Apple will get right, because of 3rd party developers.

If you get a bunch of smart app developers to focus on the 8 things most folks do every day, on the phone – check news, weather, sports, finance, email, social networks,  text messages or understand who is calling, then you can pretty much drop the need to pick up your phone by 50 – 60% of the time.

So here are the 3 unintended consequences of a successful Apple Watch launch according to me.

1. The battery life on your iPhone will “increase” since you wont “pick it up and use it as often”. Since 30% (at the low-end) and 60% (at the high end) of the stuff you use the phone for now, you can get on the watch. The battery life wont increase really, but you will charge the iPhone a lot less than every day or twice a day for heavy users.

2. Breaking news alerts, weather, sports news alerts will be more contextual and smart. So you know “just in time” instead of having to scan all of Twitter or social networks to find out what’s hot.

3. Over the longer term (5-7 years) obesity will drop among the “rich who can afford Apple watches” even further. Having a fitness tracker on your wrist that also does other things motivates you to take action.

Which brings me to Twitter.

I think of Twitter a global platform for “what’s happening as it happens” even before the media organizations get to know about it. Twitter knows first. And Twitter’s job is then to let everyone else know.

Well if you can summarize what’s happening and send it via a notification in a smart way, to all those who have the watch, then you dont have as many people posting on Twitter, or retweeting, instead you will increase the # of “consumers of the Twitter feed” even more, reducing the “producers”.

The folks that are “marginal users” of Twitter will use it even less. Why? Largely because they are in it to get information, not share as much. As much as 80% of Twitter’s users consume it but post < 10 times a month.

So, I think Twitter will become less and less relevant to them and more a “protocol” which can easily replaced by other systems.

Another loser from the Apple watch will be those that depend on Advertising on the mobile (Facebook, Twitter, Google, etc).

When you have a watch and use your mobile phone a lot less, the need to view ads on your watch do not exist.

I would short Twitter big time (I should put my money where my mouth is) because I think the Apple watch will drive its value down. I might add that Twitter may go down on its own because of other issues, but the Watch adoption will drive its irrelevance even faster.

The Market cap, revenue & profit correlations of top technology companies

Fortune has a post on the “market cap” problem for Steve Ballmer. During the period from Jan 7th 2000 to Aug 23rd 2013 here is the change in market capitalization of the top technology companies.

1. Apple – 1836.30%

2. Amazon – 222.22%

3. Google – 703.44%

4. IBM – 70.7%

Those are the winners. Now for the ones that lost in market cap.

1. Cisco – (54.13%)

2. Intel – (46%)

3. ORCL (70.21%) and

Microsoft itself is (40.46%).

That only tells you half the story.

Lets look at revenues:

1. Apple – 1861.3% increase

2. Amazon – 12118% increase

3. Google – 55389% increase

4. IBM – 18.2% increase

5. Cisco – 143.3% increase

6. Intel – 58.1% increase

7. Oracle – 266.4% increase

8. Microsoft – 222.9% increase

Here is the table.

 Profit Growth % 2000 – 2013 2000 Revenue 20013 Revenue Revenue Growth % Stock price %
Apple  3046% 7.98 B 156.51 B 1861.3 1836.30%
Google  736000% 19 m 55.39 B 55389.0 703.44%
Amazon  2948% 573.89 m 61.09 B 12118.0 222.22%
IBM  46% 88.4 B 104.5 B 18.2 70.70%
Microsoft  (45%) 22.9 73.73 B 221.9 -40.46%
Intel  284% 33.73 B 53.34 B 58.1 -46%
Cisco  73% 18.93 B 46.06 B 143.3 -54.14%
Oracle  (4%) 10.13 B 37.12 B 266.4 -70.21%

What’s the story? The revenue increase for Apple has been excellently rewarded, Google and Amazon have also been well rewarded but they have done better and been rewarded less. No clue on why IBM stock has done well despite the lower growth in revenues compared to everyone else.

Where is analytics headed in 2020? An insight gathered from 25 top #startups

The most amazing part of my job is that I get to learn from the smartest entrepreneurs in the world. I cant think of too many people who get a chance to talk to 3 entrepreneurs via video conference in California at 8 am, 2 startup founders from Singapore at 1030, have lunch with 4 amazing big data analytics company promoters in Bangalore and then wrap up the night with a conference call at 830 pm featuring a recently funded analytics company in Boston.

Most VC’s get a local perspective, Silicon Valley, Tel Aviv, Bangalore, or Beijing. I get pitched from all over the world. Most investors in the valley will tell you the best and brightest come to the valley, but I believe there’s a big shift happening. More on that later.

I wanted to share one very insightful thing I learned after 25+ detailed (over 1-2 hour) briefings with entrepreneurs who are all innovating in the analytics space.

The future of analytics is in offerings based on derived insights.

I just gathered this insight, so let me explain.

Historically the analytics space was filled with services companies. In fact  consultants would take loads of data and gather insights to help their clients with their business objectives. The best known analytics companies that dont call themselves analytics companies are Mckinsey, Bain and other management consultants. Then companies like MuSigma and others decided to “offshore” this insights service. The problem with this type of offshore services business is obvious – low margins (net of 20% and since they are people intensive, they dont scale as fast).

The purveyors of the software model of analytics are those that provided a SaaS product – names such as Cognos, Business Objects etc. Companies like Kaggle crowdsourced your analytics and there are hundreds of companies providing SaaS analytics, such as GoodData, Insights Squared, etc. The problem with this type of business is that most of these software products are “generic” hyper cubes and data warehouse / data mart models. Their margins are better than services, but still nowhere near the 80% gross margins that some industries command.

Since we all know that software is eating the world, many companies in industries such insurance, banking, finance, manufacturing are all facing a threat from new age software companies, who are re-imaging the businesses.

The next generation of analytics companies are those that take the insights gathered and create an offering in that specific area so they can benefit from the insights, instead of providing those insights to others in the industry who make more money from it.

Let me take a simple example. Global Analytics just raised $30 Million. They are an analytics company. They used to provide their insights to financial institutions by way of giving them “leads”. These leads were those customers who were worth extending credit to. An average lead in this case cost their client $30 – $100 (depending on quality).

While that in itself was a big and large market, the larger market is to extend the banking facility themselves, which means with their analytics and insights can directly offer short term cash loans to those that their analytics deems are the best. The average customer in this case will make them $500 – $5000 (depending on the size of the loan). They did this via their own offering Zebit.

Now, most founders with a background in software will say “Wait a second. what business are we in? Software or Financial Services”? That’s a good valid question.

But when you get into the “Financial Services” business there’s loads of things you can re-imagine and redo the right way with a “software frame of mind” as opposed to being a “financial services insider”.

Huge difference in revenue and margins.

That’s the future of analytics.

Using the insight gathered from the analytics to offer a product / service direct to customers and not selling the insight or analysis to existing players.

Let me give you some more examples.

Lets say you are foursquare. You have analytics and insights into where people check in, where they go, what their patterns are with respect to travel.

Would you rather sell this treasure trove of data to marketers (and face a bunch of privacy issues) or would you create an offering based on those insights yourself?

The value to a museum of information that a potential customer is near their location is possibly $2.5  (that’s quite high I imagine if the tickets are $25).

Instead if foursquare offered a virtual museum tour or a personal crowdsourced guide to the museum, then they could sell that for $10 and have 40% margin on that offering.

Imagine if you had driving habits data about car owners – how they drove, what time, how fast, how safe, etc.

Instead of selling the “best driver” data as a lead to the insurance companies, who might pay you $100 – $200 per lead, you could create your own insurance offering based on miles traveled, safety of the drive etc., changing the long standing model of one-size-fits-all car insurance.

There are lots of examples that entrepreneurs are dreaming up these days and the most audacious ones I am talking to want to upend large established industries. It is both exciting and scary at the same time.

That’s exciting. Software will truly eat the world.

The age of “speed gauging”: how entrepreneurs are changing cognitive decision making

I have been on a long road trip to meet investors and entrepreneurs abroad including, Sri Lanka, the US and Switzerland (besides many in India) over the last month. The schedule does not get any better for the next few weeks, so I am very disappointed that I am not able to write as much as I would like, but nonetheless, this is an important point that’s been brewing in my mind for the last few weeks.

Entrepreneurs the world over are changing one very important aspect of decision making – the pace and speed of it.

I spoke to over 135 investors in 15 min to 1 hour conversations (some in a group of 5-8 over dinner) over the last month to figure out that investors the world over are now under immense pressure to make decisions quickly. That was not the case a few years ago.

(P.S. I did read the PG piece on startup trends, so if he’s asking investors to move even more quickly than they are, he’s asking for a LOT, which I suspect most individuals are not ready to sign up for).

A few years ago a typical angel investor (individual, investing their own money) took 1-3 meetings and a month to make a decision to invest in a company. A venture capital investor (professional, investing other people’s money) would take longer, 3-5 meetings and at least 2 months. Then the legal paperwork and negotiations began post the “verbal commitment”.

Now it is not unusual to hear investors in the US taking 1 meeting and 60 minutes to give a verbal commitment and 15 days to funding. In India, that number is changing to 3 meetings and 45 days to funding.

Most investors have 3-5 top criteria and a subset of 5-7 sub criteria for every opportunity they evaluate. The criteria is usually entrepreneur, market, product, traction, exit potential etc. The sub criteria for market, as an example might be a) Size b) Speed of adoption c) Competitive landscape d) Pace of change in that market etc.

I am very intrigued by the sub criteria for entrepreneurs. Since I operate at the very earliest of early stages, putting money or resources when there’s just an idea, with very little or no traction, it becomes absolutely important to make sure you back the right folks.

Since I am on the plane a lot and have a new kindle I get to read a lot as well. I have been reading these books and research pieces to understand how to be a better judge of people when time is limited and the stakes are high.

a. How to read a person like a book

b. Cognitive decision making – a mathematical model

c. Thinking fast and slow

I have built a 21 criteria list for evaluating people quickly (well, quickly compared to the fact that I was not doing it at all before) and I am trying to figure out over the next year, which criteria matter and which ones dont.

Before you think this is too many criteria, let me tell you that most sophisticated investors have mentioned to me that they use between 35 and 50 verifiable and “soft” criteria” and keep tweaking their top 5. Some of these criteria can be a simple yes or no and others require you to ask specific questions. The most cultured investors, who bet lots of money have a cognitive sense of evaluating every word spoken by the entrepreneur and putting them into buckets while evaluating if the criteria they are looking for are met or not.

I am not ready to reveal the criteria since people will game the system, but I am now able to process those better. My evaluation takes now about 20-30 minutes to process each individual after I have a chance to meet them for 30 minutes. Usually I do this when I have some downtime – during commute, running, etc.

The most amazing revelation to me personally has been that nearly 30-40% of my “gut instinct” on people dont match my criteria. I used to pride my people selection based on gut feel a lot more before. Let me give you an example.

I met a really smart entrepreneur in Sri Lanka. who had thoughtful answers to nearly 7-8 very difficult questions that I had, and was articulate, concise and honest. When I went back to my evaluation checklist (which I have documented on my phone), I found that I had overlooked a few important questions and decided to talk to him the next day to ask him more questions. He stumbled on them all. Then I realized he had been asked by many folks the same 7-8 questions that I asked before, so he answered them with aplomb, but questions which he had not encountered before flustered him immensely. I dont have a problem with people not having answers to questions, but he seemed genuinely confused.

I think this field of rapid cognitive evaluation is going to see a lot more research and work being done.