When training neural networks, until enough training data is collected there will be a period when the output of the neural network is full of false positives and false negatives (aka junk)
The same could be said of the human brain (a biological neural network), unless you have access to another human brain whose output you can totally trust and rely on, there will be a prolonged period when you fumble around while struggling to gather enough data to build a mental model of the new domain.
Based on my experience, the main challenge when breaking into new domains is that no pre-trained neural networks exists. During such situations, expect a prolonged period of confusion and fumbling around. Persistence (aka brute force iteration) is probably the only thing you can fall back on.
Thankfully, totally new domains seldom exists. Whatever “new domain” you think you are trying to break into, someone else is probably either doing it right now or has already done it.
That is why I built GetData.IO. It is to help people who need data to make good decision quickly find the data they need as well as people who might already have a trained model.
The limbic and reptilian parts of the human brain have had more time to evolve. Compared to these two parts, the frontal lobe of the human brain where logic resides is a more recent phenomena. Ironically, the logical functions carried out by this part of the brain, which makes man distinct from other animals, are the ones most easily replicated by machines.
The entire human body, not just the deliberate thinking portion of it, should be considered to be a neural network. Using only the thinking portion of the human body for decision making purposes is sub-optimal. This is especially true for a human that has actively engaged in calibrating his body for a specific purpose. Prime examples are deliberate cultivation and heavy reliance on muscle memory by professional athletics, chefs, actors, music composers and detectives.
Intuitive gut feel can be considered muscle memory cultivated over time for specific functions yet expressed as formalized equations. To free up time, individuals can actively convert what they “intuitively know” into formalized equations and have the corresponding functions delegated to machines. Thereafter they could either further compound the effects of this process by building up muscle memories in other domains or sit by the beach and do nothing.
Humans will always have a role available to play in the future regardless of society’s degree of automation.
not necessary to get numbers further back than six months
stock market subjected to fractal distribution
it is possible to generate returns of up to 140% per year by trading on stocks that are moving within a range
going all in on each position each time leads to a very low Sharpe ratio
Sharpe ratio should be calculated separately for method and for SnP benchmarked against US treasury interest rates. The difference is the actual returns
On Rambo Last blood
A movie is a reflection of the culture and attitude of an age. Rambo was a very popular cultural icon during the eighties and the early nineties when memories of the Second World War and the Cold War against the communist were still very fresh in the minds of the people in America.
If you looked at the world today through the eyes of someone like Rambo, you would have been able to easily draw facts to back the narrative painted by Trump prior to being elected president.
When operating in an environment of uncertainty, a decision maker formulates multiple often competing narratives in the head that best explains majority of the facts presented. He calibrates the weightage assigned to the probability of each narrative as new pieces of data become available. He simultaneously utilizes multiple ones that are assigned high plausibility in his decision making to strive for the best possible expected outcome . It is a cognitively demanding iterative activity that goes on indefinitely.
common themes between movie and Trump’s narrative
Mexico drug cartels
Mexico prostitution rings
The world is a dark place
illegal border crossing
poor border fence
Protagonist is in his 70s
Freedom fighter who fought the communist in Vietnam and Russia
Once a project’s mission statement is defined, it becomes easy to determine when to stop further iterations.
Below are the listed of statements I periodically revisit when pursuing GetData.IO’s mission to help people make good decisions by making data gathering simple and affordable.
Hypothesis 1: People no longer need make good decisions.
Hypothesis 2: People no longer need data to make good decisions.
Hypothesis 3: People no longer find it hard to get data.
Hypothesis 4: We have exhausted all known approaches to lower the cost of data gathering to an affordable range.
Hypothesis 5: We have exhausted all viable approaches to reach people who need to make good decisions.
Hypothesis 1 and 2 are existential questions, while hypothesis 3 focuses on substitute availability. These are out of our control. The only thing we could do is monitor for changes.
Hypothesis 4 and 5 focuses on economic feasibility. These we will fully focus our efforts on. Once we eliminate all none viable options, whatever remains will be the limitations we must accept and live with.
It is useful to note the lack of any mention on funding. The underlying assumption is that every successful iteration necessarily unlocks resources from the environment which is then fed back to further the compounding process. The discipline is to minimize wastage.
Relying on external funding is like utilizing margins during day trading. While earnings get amplified, failures tend to be really spectacular. One additional drawback is that they tend to mask critical flaws in the short run leading to the commonly observed greater fool phenomena in the financial markets.
There were a total of 40 attendees and we received a total of 2 name cards – net conversion rate 5%
Social expectation during a networking session is that you can approach people to talk
Start with a clearly one liner if asked about your project
actively direct the conversation to them and spend more time listening
make sure to bring name cards
Compare and contrast small businesses and medium sized businesses
acquiring new customers
accessing to investment capital
not enough time
Medium sized business
acquiring new customers
achieving work life balance
not enough time
Sales best practices for founders n SMB sales leaders
managing leads process
Develop a concrete definition of a lead and make sure all employees understand it.
Install an effective Customer Relationship Management (CRM) Tool.
Track the source.
Distribute your leads quickly.
Nurture your leads and get your Sales team excited about every prospect.
Treat your prospects like customers.
Measure everything you do.
Hold regular meetings with your sales staff and anyone else involved in the sales process.
build out sales stages n codifying it
standardized sales process see up to 28% increase revenue
A consistent schedule: You should know when and how often you are going to be
performing your sales activities.
A strong message. You should know what you are going to say and at what point in the
process you are going to say it.
Mixed media plan. Use multiple channels to convey your message and mix it up – emails and phone calls are the most common, but perhaps it’s appropriate to reach your potential customers on a favorite social channel.
focus on boosting sales rep productivity
Make ongoing sales coaching a priority.
Advance prospects faster with Value.
Evaluate & re-evaluate sales processes.
Embrace Automation and technology.
Use Analytics to always be improving.
Observations on Sales force
UVP helps users get data from spreadsheet into systems that allows easy sharing within the sales team
Sales force comes with Gmail integration
disrupted CollaSpot’s business model
Social proofing: Video where users talk about the benefits they get using a tool
Extending existing business lines
current base – market segment with higher margins using enterprise sales acquisition strategy
new business line market segment with lower margins using self service model
USD25/user/mth package self service tier
acquisition strategy is not well defined yet
partnership with companies to value add for their customers
sales force partners with SVB to throw event to teach SVB clients how to better do sales.
Helps with SVB retention.
tap into Silicon Valley Bank’s extensive distribution network
Alessandro Chesser, VP of Sales Carta
Not investing early enough in sales operations.
CRM system very important. Cleaning up the mess later is going to be a headache
Dealing with duplicate accounts are pretty painful
Don’t over engineer sales process up front.
Enterprise sales versus organic adoption
Organic signup forces u to put your whole pitch online for easy copying. A sales rep can sell the vision to extract higher margins. Deliberate back and forth. Use the demo form to generate leads for sales rep
Getting the enterprise sales engine started
The more data you have the better your decisions. Create baseline and iterate over it over time.
Marc Benoff on Closing. When he joined eshare there was no product. His job was to get potential customers excited about the future of their own company. The vision and the pitch is so important. Get their feedbacks.
Don’t sell to far ahead and can’t deliver. Need to spend time with engineers to know what can be delivered. Always keep in mind a 1 to 2 months implementation cycle. Only sell ahead when is early stage company but don’t over promise. If failed to deliver, will create bad PR
Always iterate on the vision of the company and the pitch to figure out what resonates with your potential buyers
Early stage startups can utilized VC to generate inbound leads. The more money raised the more inbound happens with PR that follow each fund raising events
Scaling the sales organization
Make sure one person can bring in 100K ARR before replicate and scale up sales process
When is Sales cycle replicable? Use revenue as signal. Driving 100K ARR per month is a good signal. Gut feeling. When deals leads are starting to slip through the cracks
Charge via ACH instead of credit cards. ACH is more scalable, since the latter tends to expire.
Structuring and managing the sales team
For smooth transition from inbound sales to outbound sales first create demand from SDR first before hire sales team.
Not all sales people are comfortable with generating sales leads
Extremely important to make that SDR hire early. Utilize tools like OutReach.IO
SDR – handle email marketing and phone calls. Uses pitchbook. Scrape Startup names and email addresses. Thousands of emails a day to generate leads.
Ensure at least sales development representatives hit at least 60% their quota. If hitting below, it means u have over hired. It’ll create very bad culture like sales leads stealing if not hitting above quota.
Have sales people prioritize and focus on closing and not be the jack of all trades.
Promote Successful SDR to become sales reps. They will be well positioned for success.
Mapping the hiring process for the sales leader.
The mistake is hiring really experienced and expensive people who are not willing to roll out his sleeves. You need someone who is really willing to get his hands dirty to go out to close sales. Industry experience is important, will ensure sales leader motivation level. Since he knows what is broken
Balancing the functions of marketing and sales
Early in startup marketing and sales goes hand in hand. How to ensure no stepping on each other’s toes?
Marketing organization should have good process too
Need to balance load of marketing and sales organization. Make sure invest more heavily in marketing to generate more leads than sales can handle to ensure good culture.
Pricing is important. Need to make sure not too cheap. Perceived value is very important. Need to be more expensive than competition and explain the value clearly to targeted subset of customers.
Learnings from Loominance
UVP to attract customers via word of mouth: They need us when they are drowning in data.
Account based marketing
Scrap company websites, identify and recommend similar companies as sales leads to clients
The decision maker and the executor operates with different cognitive models.
The decision maker
The decision maker operates within an environment of incomplete information. He takes in conflicting signals from the environment to figure out the underlying Markov Chain that holds his environment altogether.
A successful decision marker minimizes the price paid for knowledge of each section of the Markov Chain.
Each success unlocks more resources that could be redeployed to this activity.
Inner fortitude is the primary trait required to navigate the high failure rate inherent to the nature of this activity.
Perpetuate accumulation of mental models from different domains reduces taken to form judgement while increasing overall batting average.
The executor operates within an environment that has been made predictable by the activities of the decision maker.
He is predisposed towards systems, structures and processes. A successful executor measures and optimizes with the goal of increasing yield from a well defined process while minimizing costs.
Disciplined consistency is the primary trait required to ensure continued excellence in this repetitive activity.
A strong decision maker profile with a weak executor profile will feel repressed/bored when forced to operate as an executor in a structured environment.
An strong executor profile with a weak decision maker profile will feel overwhelmed when required to operate within an unstructured environment.