Do founders need forecasts/predictions?
Hi 👋
I need your help founders so i'll keep this as brief as possible.
I have a finance background with 10+ years experience, now data scientist with specialization in forecasting and predictions.
At first I wanted to forecast markets (stocks, crypto etc) but couldn't find the right customers. Now, I'm trying to find the best initial customer for my startup Cuanto. Do founders need highly accurate forecasts using AI and Machine learning?
If yes, for what?
Alternatively, if you have other suggestions of how I can best use my skillsets to serve founders, CEOs and CFOs please let me know.
It'd be great to know the future, but I just don't trust many forecasts.
Love the feedback, what would help you try a forecasting tool?
If you can consistently forecast with >51% accuracy in the financial markets (such as crypto, stocks, etc), getting extremely rich should be relatively easy. Getting customers would be largely irrelevant. The challenge is getting an accurate forecast, not getting customers. Everyone would pay a fortune to know tomorrow's stock price, today.
There's far less economic "value" add in sales/revenue forecasts, then in future asset price forecasts. They might be good for pitch deck modeling, tax planning, etc. That's a $5-500/mo SMB SaaS product. An accurate financial market forecast model is a 50K+ a year alternative data product that could be sold to large institutions.
Thanks Mark, appreciate the feedback!
So there has been success on the market forecasts in the most unpredictable assets - stocks. I was day trading for a few weeks with daily forecasts and they were 80%+ accurate still. I was also using insights to see exactly what stock signals (over 30) were signaled for each stock and how effective they were.
The forecast has been super accurate on predicting the further market crash. *this is without any updates.
I'm finding your comment super helpful, can you elaborate more on this please? 'An accurate financial market forecast model is a 50K+ a year alternative data product that could be sold to large institutions.' - predicting foreign exchange, overall economies etc?
this was the last article I wrote for Cuanto which was forecasting QQQ (Nasdaq 100) back on 8-29-2022: https://us14.admin.mailchimp.com/reports/show?id=9046605
Sure. If you can predict future asset prices with statistical reliability, hedge funds, institutions, etc may pay for that. They'll do so to enhance their prop trading strategy. It's often called, "alternative data". Certainly if you could predict FX (one of the most unpredictable markets) there would be huge value in that. There's less value in predicting overall economies, as there's just not as many ways to directly monetize that. Still, it could be valuable for certain groups, who might be willing to pay for such a data feed.
50K for a feed is just a rough guesstimate, but consider that a Bloomberg terminal costs about half that - and how many thousands of those are out there? Similar market.
By the way, the article you linked isn't publicly viewable.
I don't quite understand your proposition.
Forecasting is always useful, but it's expensive. I am very early-stage and I know money/time is better spent on acquiring than forecasting right now.
Let's assume I have $5k MRR - do I need sales forecast? I can pretty easily run a calculation on churn and acquisition and then see what my forecast looks like.
Similarly for expenses, might be a little more complex as there might be stair-stepping in some expenses.
Do you propose doing that style of forecasting?
This is immediately useful, and I think there is a market there. But I think it's a harder sell as founders likely know how to do limited forecasting. How would you ensure your insights are worth more than your cost?
Or do you propose more broad forecasting?
Leroy thanks so much for the response and feedback!
In my last role I was the lead for forecasting total MRR, territory MRR, fp&a spend, price outlooks and benefits expenses. To your point, this was a 2.5+ billion in revenue company so there was a strong need for this, but wasn't sure if startup founders could use these forecasts. I think FP&A and small to mid-sized business would benefit from this more.
My proposition would be save $160K+ annually and get 80%+ accurate forecasting with just a click. So if the SMB has $1M ARR then they can get value from highly accurate models without hiring expensive Sr Financial Analysts too early.
If you truly expect you can convert 16% of revenue from cost to earnings, then yes, there will definitely be demand.
But I am not convinced about 'just a click'. Surely you would need to see data of product stock/orders/return etc.
I am also not sure if you expect you can save that amount with physical or digital businesses. Do you think you can cater to both?
meaning 16% cost avoidance accounted as new revenue? this is certainly possible. This was one of our last projects where we were doing more data analytics to find inefficiencies in spend and sharing insights, resulting in hours saved or $ saved.
For insights, data exploration would be best. But for forecasts we were able to do 'just a click' with my auto-tuning algo. This is where some feedback from the user would help tune the model would help the AI tune better.
yes, physical and digital both can be catered. The more columns (stock/orders/return etc) it's trained on the better it can perform. I got some really interesting insights on some of the data showing how the end of life products were throwing the forecast off, so it first predicted them going to 0 then got better results.
Thanks to those who are also voting! 🙌
33% voted forecast for 'Other' so far, what is the other data that you'd like forecasted? Churn?