How to talk to your user
We are an early stage startup that focuses on Q&A for fans and creators. The goal for the startup is to get every single questions on earth answered. For that, we’ve developed internal algorithm to find the relevant fans for each question based on their past posts comments using NLP and AI. Ultimately, we hope that through Q&A channel, we can store and index more human knowledge than ever.
Because all of the above is more or less a dream, (technical part is not the hardest to solve), the biggest question is whether this is actually a real problem that irks my users to a great extend. Given that we are a Q&A for fans and creators, we’ve started a pilot project with a celebrity chef Joshua Weissman (aka papa). The goal at initial stage is to find product market fit by talking to my users, mainly papa’s fans.
How do I talk to my users? I’ve checked out YC curriculum and also read over the books “the mom test”. Not that I am an expert in any of it, but I took some notes and here are some of my takeaway, mainly for myself to remember: (the book actually covers a wide range of startup, but this is some take away for early stage, consumer company focusing on building communities):
(1) Never start a conversation (early stage) by pitching your ideas — aka the mom test
When you first tell user about your idea, most of them are nice, like your mom who wouldn’t want to crash your dream, and they would say something like “oh it sounds interesting” and compliment you along the way because you are in a sales-mode and people know when you are trying to sell. Or worse, a user would argue with you about the validity of your idea and then you try even harder to sell it to him. Eventually to end a conversation, your user says something like “now I get it “ and you would even feel great about it. The reality is that NOBODY can tell you whether it is a good or bad idea except the users who are using it. To see them using it in a real-world setting trumps the best investors’ compliments. In general, in early stage while talking to consumer, you are either selling or learning. It is usually difficult to first learn and then to sell, users are likely to feel that the conversation is not genuous.
(2) Instead, try to have a genuine and very specific conversation about the problem that you want to address
In my case, I would ask them when was the last time that they cooked one of Papa’s or any youtube chef recipes. Once I can hold on to a very specific dish, I will ask the users how he prepared to cook, whether he faced any difficulties, whether the dish was successful etc etc. The goal is almost to invite them to tell you the specific reality story, not just during the cooking process, but the path of the beginning from how they even get the inspiration of cook that dish to the end, including how it turned out, whether they tried in second time? etc etc. In early stage probing, concrete data points are not actual data that you collect on web analytics, but what users have actually done specifically in real life, after talking to a few of them, you might be able to find patterns and then form your hypothesis. Questions like “would you, could you, do you see yourselve ever..” are bad questions because they are hypothetical and users will almost never know what they might actually do in reality. However, they can be stepping stone questions, “oh so what is the last time that you actually did this…”. The key is always to bring to a specific instance and make it a concrete data points
(3) Form some hypothesis and 3 big questions
After talking to some amount of your potential users, you’d likely to see or hypothesize what problems that your product is likely to solve. For us, our current thinking is to target users at two stages of cooking (1) Pre-cooking for questions regarding substitutions and modification for dietary restrictions. (2) Post-cooking when a dish didnt go well and the user would like to know what happened there. As I continue the conversation with my users, I would try to ask questions about those areas. “ Do you have any dietary restrictions? Did you end up getting all the materials that you need for the dish”, or “Have all the dishes turned out great, were there any dishes that didn’t go well” and then invite them to tell me the rest of the stories. In general, while I would want to listen to the stories to see if I am missing something, I also like to guild the users to points that I want to learn to see if it fits into my hypothesis. The rule seems to be prepared and try to get 3 of your most pressing questions answered before reaching out.
(3) How to set up the talk?
I think a goal of mine is try to talk to users every week. A big mistake that I’ve made is to make it super formal. I’d ask “oh I am working on this Q&A startup for you guys (sin no.1!!), can I use 10 mins of your times, quickly talking to you and asking you some questions (sin no.3!!)”
Sine normally customer survey for early stage startups are super short, there is no need to set up a meeting and make it all former. This especially makes gen-z extremely uncomfortable. (It’s unclear how often they actually call people up, or ever). Instead, if I want to ask a question, I just DM them and ask the question. For example, a user shows a great dish on discord, or I need to do to start a conversation is simply by typing “hey! this looks awsome! Was it a difficult dish to cook?” and start a virtual chat about it. There are only 1 formal way of talking to your users, but about a dozen way to talk to them informally.
(4) A disinterested user is good news.
The goal of any conversation is to learn. As a result, the worst kind of conversations are the fluff one where you get no concrete data points but a lot of compliments. A dis-interested users, or a user who would never use your product is a good conversation because you learn: (1) He’s definitively not your customer at this point at least, (2) Maybe, just maybe that this is not as real of a problem as you think. But I guess you would have to talk to good amount for this learning point. Getting rejected is hard, but it is far worse to find out after you spend all your resource and manpower building it.
(5) Segment and then segment again
Having a product that is too generic is never a good thing for an early stage. The issue is that if the product is for everyone, then it is not super good for anyone. Another problem is that you are likely ended up talking to 1 customer per each group for 20 groups, rather than 20 customers within 1 groups. Your learning on your former set would be much shadower than your learning on the second group.
There is always more detailed segmentation that you can drill into. We started out thinking of youtuber and their fans => cooking youtuber and their fans => Joshua Weissman and their fans. As we talk to more and more of his fans, we realize that his fans can also be into two categories: (1) The one that casually cooks (2) The very serious cook that studies and researches everything. I am sure that as we talk to more people, we would segment his fans into even more specific categories.
(6) How to take notes?
A user interview without notes or with all the notes are akin to an interview that never happens. Memory fades and the purpose of user interviews is to share what you learn with the team. The book recommend an effective way to record the conversation in a concise way without losing its authenticity: write down the direct quote from the user and then annotate users emotions to it. In this way, one preserves the originality and the context. A good idea is to categorize those statements and then organize them into a spreadsheet for future reference.
Other cool things that I learned:
(1) How to cold-email people (might be useful for me later):
First, mentioning that I am an entrepreneur that is trying to solve this problem in a high-level sense. Second, talking about the stage where you are at trying to solve the problem (super early stage). Third, talking about the weakness and the specific issue that you are facing. The potential reader at this point is expecting a sales-pitch, but by showing your vulnerability, you are showing that it is not a sales call. Fourth, compliment them and tell them why they are the best person to talk to about this issue. Fifth, do the specific ask.. 20 mins of your time etc etc.
(2) There are only good and bad meeting. There is nothing in between
A good meeting is a meeting that has a specific next step and commitment. (This is probably more applicable when you start to pitch your product). A compliment doesn’t count as a good meeting without a specific next step. I would also add that a meeting with a goal of learning is a good meeting when there are concrete data points being learned (even if the data point is that the user is never gonna use the product!). If a meeting doesn’t fulfill its goal and has no follow up, it is by definition a bad meeting. There is simply nothing in between.
