For the last 8 months, I have been building a SaaS for b2b. Basically, it is a service to analyze a user’s customer list and give insights and generate prospects for conversions.
I am a developer who specializes in backend development and machine learning algorithms.
I have built the entire service as an API. It can manage users, upload and manage customer lists, run a basic analysis (using analytical means as well as ML), generate insight reports on the customers to provide demographic/behavioral/psychographic information, generate new prospects using ML based on the profiles, and a host of utility functionality.
I am currently working on the MVP front end.
I am aiming to provide these services to small and medium businesses as they would otherwise never have access to these. During all this I talked to many friends and businesses to try and validate the idea and though many like the idea most have difficulty understanding all parts of it. Those that do understand are willing to pay but how much they are willing to pay is all over the place and most do not understand using all the services continuously. However, these could be aberrant.
As I get further and further along, I am having a lot of trepidations. For one I am not a business-minded person, so it is difficult to figure out exactly how to market it. I also don’t know how to get investors or funding. For another, I am losing faith in it being profitable as it seems many businesses would not pay a recuring fee as they might only want to do it occasionally and the entire cost of the endeavor is high.
For a project like this it would be about $350k to $500k per year and I am not sure if I can even find funding or investors. That cost includes data acquisition (this system needs a very large accurate dataset that is updated regularly), hosting for the service as well as the ML processing power, development costs, marketing, and possibly salespeople.
So, it comes down to this.
Do I try and pivot and offer only a very small portion of this service, say just the first analysis part? The advantage of that is the cost savings as I could build the MVP on my own and it does not require the large dataset. The disadvantage is that I am not sure if people would subscribe to the service if that were all that was offered.
Or do I try other options?
Move on to another idea?
Try to get a partner or partners and give them major equity (but who would do that)?
Just give it away and have someone else finish and work on it? (I fervently believe this is a good product that fills a niche that is missing – giving these services for small businesses)
Any advice would be greatly appreciated.
OP, we are also an AI app, and we use lots of GPU for our own model training as well as using 3rd party ML APIs and LLM models. Try applying for Google for startup, MSFT for startup programs. We got around 350k in credits (in total from both), and that has been very helpful.
They might require you to have some angel investor or a product, but it is worth asking.
Good luck
Thank you, I was not aware of that. That is very helpful.
Really interesting product! When you say it needs a really large data set, how big would you say is the minimum number of rows/sample size?
Can you elaborate on why it would need to be 300k+? Is that all infrastructure cost? That’s a very large ask even for established B2B companies to be making.
It needs a consumer file with demographic, behavioral, and other data. The data set I work with on another project is 600 million records with about 400 columns of data. It costs between $10k-$15k /mo depending on the fields you choose and update frequency. Hosting, managing, and updating that kind of data set also is costly.
How does your target custom gain that data to begin with? Im trying to conceptualize the customer journey and business context. I can see the value in making sense of big data, but it assumes the customer has the data to begin with.
My user would have their own customers (say they are an online shop where their customer has bought something from them) and would have some information about them, say name and postal address. They would upload their customer information into my site and the data we have would be used to generate insights or run ML modeling.
Sounds like you started backwards with a service that you hope somebody will need. Take a pause and reach out to your ideal customers first (identify them first) and start talking face to face with them. A market that doesn’t exist doesn’t care how clever you are. Right now, you need to learn from your customers directly and pivot often as needed. Check out the book “lean customer development”. Stop spending so much time obsessing over delivering a perfect product for nobody. Figure out how to deliver value. Take a pause before you spend more money and time.