market-researchI know almost nothing about startups and never know how to judge whether my ideas actually merit a serious attempt at a startup.
I have made several efforts in the past to educate myself about how startups work. However, few resources I have consulted have offered (to me) a convincing method for how to measure/gauge/estimate the growth potential of a startup business. Let me illustrate what I mean through an example.
Suppose I invent an unusual and desirable way of making pizza. Let’s say I have managed to invent a pill that, when water is added, immediately morphs into a fully formed, deep dish, Chicago-style pizza (complete with toppings and already in a box lol). How do I judge whether this product is marketable? Let’s suppose I take my pizza pill and sell it around my neighborhood or my city. I thus get a small sample size of what potential my pizza pill has for that cluster of customers. Suppose I also have extracted data on them because I have them use some sort of app or online interface that permits me to monitor certain buying behaviors. But of what value is that data? Can I take data about a local region or small cluster of people and immediately extrapolate to make claims about the growth potential of my business to a much larger sample size?
Perhaps my background as a mathematician and physicist makes me take mathematical claims slightly more literally than others. When I make a claim about what a number represents, I want testable ways of demonstrating that the number means what I say it does. If I am measuring temperature, I don’t want to estimate my refrigerator is 50°F and have it be 100°F. My milk, eggs, yogurt, meat — all of it will spoil. In other words, that means all my resources go to waste. Similarly, if I make an estimate of the number of customers I expect to buy my product, I want to know (within a margin of error) that estimate will actually be accurate. I want to know there is a very high probability (not a high hope) that the resources I invest in my venture are going to yield profitable returns. On the flip side, I want to be able to identify when an opportunity won’t be profitable so I can save my assets for ones that will be.
To my knowledge, no such method exists. Perhaps it is too idealistic. If that’s true, what are the alternatives? Having failed to find info on this myself, I have turned to SE in the hopes someone has ideas about how startups and entrepreneurs approach this problem.
Pre-orders are the only valid predictor in my experience.
Statistical data obtained using models offer zero predictive value. We’re discussing humans here, with desires, moods, etc.; they’re not predictable automatons.
Another way of doing the same study is: ask friends and their buddies if they’d buy whatever you’re thinking about – and they all promise to buy it.
It also yields zero predictive value. Because friends and their buddies will say they’ll buy it as a favor to you. Some might even actually buy it. Real clients won’t.
Here’s yet another: ask a sample of your target audience if they’d buy it – and they all promise to buy it.
That – yes – also yields zero predictive value. (I’ve seen this first-hand, and it was not pretty.) Because, the only one who commits to anything when a promise is made is the person who receives it.
Here’s one that works: ask a sample of your target audience if they’d buy it right now, and collect their signed commitment to do so – to be delivered later.
I realise I’m revisiting an old question, but thought I’d throw in some alternate ideas that may give some food for thought. Especially given the amount of money being thrown at every sort of business there is by investors and VCs at the moment. My background is in data analytics/data science, so I have a solid basis in stats and I run my own business, which I’ve done both very poor and some very good forecasts upon.
I’m making the assumption that we are dealing with a B2C business, selling a consumer product/service.
Starting points which give no basis for good forecasting
Starting points which give little, but poor basis for forecasting
Starting points which give some basis for good forecasting
But still even with information from the latter category, you are bound to make large assumptions to extrapolate a forecast.
This is why online businesses are so popular for VC investment, they can be up and running with a certain amount of effort/cost/time and be marketed through appropriate channels to the target audience. It is very easy to run a “proof of concept” and then scale on the back of investment.
For bricks and mortar businesses, this is more difficult to get a business running and then scaled. Having one shop front with associated overheads, staff, marketing is a challenge to prove an initial venture, but to prove it in a number of cities or markets becomes logistically exceedingly challenging. E.g. Starting a “Hippie Style Coffee Shop-Bookstore” in the Soho of London (UK) would almost certainly work, but transplant the same idea to Ipswich in Suffolk would likely not. So then you cant get the advantage of economies of scale to bring buying efficiencies to your business.
I would be collecting every bit of data on the sales of your product/service during the time that you are in “Proof of Concept” stage. Get to know your customers. If you are online it is easier and can be made more automated, but bricks and mortar business need the same information to thrive. Perform surveys (online or face-to-face), collect data from the app store that you can. Give away free products in exchange for customer data. Then you can then segment and understand your customers and make much better predictions about who your customers actually are and why they buy from you.
Once the market has been established by actual sales, then start scaling, continue to gather more data, enrich your forecast models.
Another thing to note, is that any forecast model you make, you should measure the goodness of the model in retrospect, so that you can improve the model, or in some cases completely discard it!
All content is licensed under CC BY-SA 3.0.