Selling Use Cases vs Selling Solutions to Problems
I recently came across a thought-provoking tweet.
My interpretation is that transformational software tools enable many new use cases that lead to new categories of applications, leading to more future value creation than solving a specific problem. Having a clear problem in mind guides you to building helpful something but if you’re too focused on the specific problem you run the risk of adding only incremental value.
At my company NLP Labs I think about this tradeoff a lot. Will we be an API company or an end-to-end product company for a specific niche like customer service? Right now we’re trying both as parallel bets and will react to what the market tells us.
In the case of the API, I’m selling to use cases that involve extracting information out of documents. My personal feeling is that the API business would power many processes and products down the road as Erik alludes to in his tweet. I believe total value creation and impact would be much larger not to mention the optionality of different markets.
The con of selling use cases is that there’s a customer education piece involved where you have to tie the API to improving some bottom-line business or product metric. If you’re not thinking about NLP 24/7 like I am, it may not be obvious how things like semantic search and automated question answering help your business.
Whereas for solving a specific problem I currently have a Slack app and soon will have a chrome extension for customer service teams that uses semantic search and other NLP technologies to help them resolve support cases faster. Whether the product uses NLP or not is irrelevant, customer support managers only care about the bottom line metrics. The sales process involves little to no customer education because the metrics we’re targeting are tangible and “hair on fire”.
The con to focusing on just customer support teams and solving specific problems is it can lead to a chain of decisions that only prove useful for customer support teams and we learn less about other applications of NLP. We’d also have to focus on non-NLP product tasks to build an overall great product but these wouldn’t be our bread and butter. If we go down the CS rabbit hole and end up not selling it that’s a large risk and big opportunity cost.
I suspect this is a common dilemma in software entrepreneurship but now that I’m at the helm it feels a lot more real. If you’ve gone through this or a similar dilemma or just want to chat about anything NLP or entrepreneurship-related I would love nothing more than to hear your thoughts.