After A Number Of Crazy Conversations With Potential Entrepreneurs About Complicated Business Models With Lots Of Moving Parts As Well As Making A Few Mistakes Of My Own I Felt Compelled To Write This post….

OCCAMS RAZOR & SIMPLICITY Horst.Burkhardt CC tldr: cut features don’t add them Note: You have my permission to republish this article, check the notes at the bottom. Occams Razor Occams Razor is a theory ascribed to a 1300th century Philosopher which states “Non sunt multiplicanda entia sine necessitate” (Entities should not be multiplied unnecessarily) or paraphrased “When two alternatives produce the same result always choose the simplest alternative”. That is its always better to aim for simplicity instead of complexity. Too often I see experienced and new entrepreneurs (myself included) making things far more convoluted than they have to be, failing to take the simple path and I wonder why. Unnecessary Business Model Complexity I recently had repeated discussions with a fantastic technologist I admire who has struggled to translate their super cool machine learning work into a viable product or business model. See if you can guess whats wrong, New cloud based machine learning technology+ Needs to be integrated with Hardware + Needs its own App (iOS and Android) + Content Licensing issues with 10,000s of owners Multiple Parties wanting to cash in on advertising 4 other parties in the commercial negotiation all with compelling reason to kill this, Tenuous claims to improve advertising revenue Its not clear who pays and who is the customer What could go wrong? I see this a lot with inventors and inventions as well. In a previous job I had to review a lot of inventions and patents (over 500 in last few years), the moment I see multiple parties in a complicated relationship where its not clear who pays or who’s problem it is, we run a mile. If you can’t explain clearly who really needs the solution, who is the customer and how you get paid when you are in idea stage, its not going to get better when its a startup. Technology in Search of a Problem You might recall my approach to startups and inventions is to make a decision whether I believe the problem is real and who really cares about it. I call this the “Who gives a fuck?” test. If you can’t convince me that someone really cares about this problem I struggle to get interested in the rest of the pitch. These complicated business models occur because the would be entrepreneur has developed an interesting (to them) technology solution first and then set out in search of a problem. We see this repeatedly with University research. To make the technology look viable (and attract grant/equity funding), would-be entrepreneurs invent a customer usage situation to try to force fit the technology into a problem. Often the connection to reality is tenuous but they can’t see it, no one including themselves can think of a situation where a user would find the capability compelling but it doesn’t seem to bother them. This happens with every tech wave (and I have now seen a few). IoT, Big Data & Analytics have been the must have hot pitch buzzwords for the last few years (with Machine Learning catching up fast). With every man and his dog making an IoT device or solution (yes Pets are big in IoT) I hear a lot of talk of selling advanced analytics data to some imaginary customer. When pressed no one can describe this imaginary customer. Most of these business models are predicated on large scale adoption of hardware/big data solutions with the payoff coming when they sell the data. Over the last 3 years I have had too many would be startups try to convince me that customers will actually pay any attention to their convoluted analytics, let alone pay for them. As an example I purchased a Fitbit and a Jawbone over a year ago and wore them both for months. The only thing I paid attention to was the number of steps. Same with Apple HealthKit, the only thing I check is distance covered and steps. I might want to see data for a few weeks, but I really can’t deal with more detail than that. Most people buy products and apps for a few key capabilities, dozens of additional features and complexity often hurt your solution not enhance it. At the start your product and business only has to do one very important thing really well (and thats up to you to decide). Most people want the key measurement at a glance or key functionality executed well, complexity doesn’t help solve their problem. I know Fitbit has a stack of detailed measurements and other health related parameters you can input, but I lack the attention span to spend an hour looking at detailed stats for my wearables and frankly it does’t help, the time is better spent walking. I see entrepreneurs (you know who you are) adding so much complexity to really simple yet compelling products and as a result a great opportunity is lost as the user is confused, unsure what they should pay attention to. Im mindful that the Fitbit is a personal device and that some IoT business cases may have more need for data, but the frequently cited business case of “selling data” to someone else once they have critical mass in the market is unlikely in most cases. Adding complexity to the product doesn’t help and creating a complex business model to fit an imagined problem is nuts. An Example of Unnecessary Complexity There is a SaaS product called Intercom.io, it is a pretty good product for helping customers on your website or app. I admire their founders after hearing them on podcasts and read their books on software product management (which I recommend), I have mentally purchased the product, however their complicated pricing model stumps me. They have 5 different packages with pro and standard for each package combined with a per user count, but I cannot for the life of me work out which one(s) I need, after looking at all the packages with various overlaps I think I probably need them all (as a startup you end up doing it all), but I feel like they might have easily helped me and said you need it all. To quote the CEO in a recent podcast “The customer rarely buys what the company thinks they sell”. There are infinite possible pricing dimensions to this pricing model and it makes me feel dumb because I don’t know what to do and everytime I want to use a useful feature they push me to upgrade. It would have been much better if they just made it 1–2 packages with a per user count and I would probably be a paying customer by now. Far be it for me to criticise a company which is arguably pretty successful, but it’s hard to imagine Im the only one who feels this way about their pricing model and it must be costing them new customers. (Update I am now using the product, but Im still at a loss to the dozens of variations and have no idea if I have the right product and it spoils what is already a lenghty activation process with multiple parties needed to get it implemented). Take features away don’t add them I recently interviewed Allen Liao from Tzukuri.com who launched the coolest smart sunglasses (they are equiped with a small microprocessor and bluetooth and are virtually unlosable). When I asked him how an undergraduate Electrical Engineer with no design background could build such a beautiful product with such a classy design aesthetic across web, marketing and the actual product, he replied; The secret is to put less on the page, take stuff away don’t add it. A few points about complexity A validated user problem stated simply, cuts out a lot of crap Simple beats complex most times Most people can execute on simple, they just can’t work out what simple should be. Teams have a tendency to introduce more complexity, they struggle to say no to features. Simple is harder to define but is easier to build and implement. Simple is easier to sell to everyone, investors, team and most importantly customers. Complexity prevents you acquiring customers, if you have to personally explain a complex business model for a customer its much harder to convert customers. Customers have bullshit detectors as well and subconsciously vote with their feet if they get cognitive overload. If multiple planets have to align for every customer for your business model to work, you probably won’t get paid. There is nothing wrong with a simple business model, i.e. we sell stuff or we charge for a service. Don’t Believe me? Listen to the Billionaires Warren Buffet and Charlie Munger sum it up best, Simplicity has a way of improving performance through enabling us to better understand what we are doing. If something is too hard we move onto other things. People underrate the importance of a few simple big ideas — the chief lesson is that a few big ideas really work. If it’s any consolation I am not immune to this problem. In my spare time I have been working on a sideproject called Newscloud. Newscloud.io is the API for News. Essentially Newscloud delivers a real time news and content stream, tailored to topics you are interested in. We check 160,000 curated websites every hour and sort the fresh stories into 5 million topics. While building Newscloud.io we have struggled with complexity, not knowing when to add or cut features (many of which the user can’t see). We built it for Social Media Managers, Journalists and Bloggers who want to find and share the latest stories and perhaps later for financial traders and algo trading platforms who want to trade off real time news. Dyson our trained monkey checks your Twitter feed and works out what topics you like to Tweet about and then gives you a personalised story feed that matches your favourite topics. The trouble with building a real time personalised search engine is that there is a lot of tech that the average Google user takes for granted but never sees. Language detection, spam detection, topic classification, computer generated summaries, semantic search and the complexity of algorithmically deciding which sites are great and which are spam, as well as the logistical issues of crawling 160,000 sites every hour without pissing them off. So many complex systems to reduce the firehose that is the internet with 10's of millions of new stories a day into a simple but comprehensive feed of stories that are of interest to an individual user in English without spam or porn. Simple right? The decisions we struggle with is how much of that complexity do we let the user see and how much do we hide? Although we want to let the user know how complex this is and how hard we are working on their behalf and how many millions of stories we check each week we have decided that users should be completely oblivious to the tech unless they choose to go digging. Specifically we are removing a lot of the complex features we felt we needed to show the user like trust rankings (it shouldn’t be in the index unless we trust it) and sentiment (no one really cares) and other things that cause clutter. Instead of showing 1000's of irrelevant stories when users land we are going to only show the hot stories and require the user log in because thats the only way we can give them a customised experience, otherwise its 200–300k stories a day that are not relevent to the user. The aim is that the user doesn’t see anything which is not 100% relevant to their interests. Because it’s so hard to vote features off the island once they are in the product we will take the approach that every single feature must be knowingly voted onto the island. Frankly we are nowhere near product market fit, not even close, there is a long way to go but taking some of my own advice is the first step forward :). Newscloud.io is now in beta, if you are a social media manager, journalist or blogger we would love your feedback or suggestions, login with Twitter and get your own customised news feed. What can you do? Fight to keep new features out of your product unless you vote them in. Look at existing features, customer segments and functionality and keep cutting until you have removed everything you can. Rewrite your website copy, articles and ads until the simplest compelling message remains. I could write another 10 points, but I think thats simple enough :). Republish This Article Recently I have had an expert commentator steal my two of my articles (from different sites) and remix and republish them on commercial sites with absolutely no acknowledgement (yep, unbelievable). In response I have decided to start granting permission to legitimate publishers to republish my articles. This article has been given a Creative Commons Non Derivative licence. Basically anyone is welcome to republish this post exactly as it is without changes with full attribution and a link to my Twitter account @mikenicholls88 Thanks Mike Nicholls PS Love to hear your personal stories of simplicity and complexity screwups