Lesson learned: I've overcomplicated things
by Lizzy TrelstadI started my company with the idea that I could use chemistry to read personal care product ingredient lists for people, and then report back to them what was best for them. Emphasis on "for." I came up with the idea of "chemist curation," and this chatbot was the first tech tool made to help people discover their next favorite products.
I've been closely watching the conversations people have with the Beaker Bot. People seem to like the idea of being matched with a product, but are getting frustrated that they didn't know why they got what they got. I thought them using the chatbot was the same as them opting in to trusting a chemist, and that that was sufficient. I thought I wouldn't need to explain the chemistry, I could just use it.
I was excited by the idea that I could automate and scale a chemist's advice, without worrying the consumer about the nitty gritty chemistry. From feedback and conversations that broke the chatbot flow, I've learned that even if people don't know the nitty gritty, they still want to see it. They don't want a chemist telling them what's best for them, they want a chemist that's telling them what's best for them, to them! My product isn't the end recommendation, it's the process by which I arrive at the recommendation.
With that known, so much of the friction I've been feeling as a non-tech solo founder has disappeared. I don't have to build my expertise into tech, I can simply use my expertise.
I've now decided to release myself from a need to create a "tech product," and have decided that for the next 3 months I will produce as much content as possible. I'm going to do the things that don't scale so that I can build a large and loyal audience. By the end of the year I'm hoping to have an audience that can tell me what they want, instead of betting 3 months' time on creating "tech tools" that get at what I think people want.
I think it's possible to be an innovator while still listening to the market. That's my new hypothesis at least; I used to think that being an innovator meant being ahead of the market. But that's a sure way to make a faulty business model.
Just because I don't yet know exactly how to help my audience, and therefore what scalable product to build for them, doesn't mean I'm not the domain expert that I am. This was an important lesson to learn.
Does this make sense? Or am I just rationalizing my failure to hit upon a scalable product? Any and all pro-tips and advice are very much welcome.
-
3
In my own past, I have put way too much emphasis on building something scalable right out of the gates. I think that doing things "manually" will give you a ton of insights and allow you to navigate your product development until it gets to the point where it really needs to scale. I don't think planning to scale is wrong, but it should be an ever-changing strategy that evolves as you learn from your customers and market.
-
1
Great to hear that this "manual" approach is validated by ppl who've been through the wringer before :)
-
-
2
Makes sense! Keep going and never give up.
-
1
💪💪💪 thanks!!
-
-
1
Hey Lizzy,
Really like the idea, and I think it fits a user need. Good potential for affiliate marketing and referrals, and plenty of opportunity for evergreen content to drive users to your site.
One issue I have is the chatbot -- why did you choose a chatbot ?
Instead I would opt for something like a typeform survey.
Some reasons:
I don't like waiting for an automated bot to type me a message.
The chatbot doesn't allow for good design. The long message content is unnatural to read and would be much better as a button with a large description. And you could hide extra content in a typeform and let the user decide if they need more information for each question.
For each question, i get two bubbles, and have to move my eyes, which is unnatural.
I really like your idea, and I think to give it a fair shot it needs to be a typeform or similar.
Something to consider:
Things like "Go meow" -- cute, but adds cognitive load. I didn't get it at first. Did you have users use your product while you watched? Did they pause there?
A lot of the answers I have for questions don't fit within the boxes you offer, and that also produces cognitive stress, meaning users are more likely to drop out.
Also with a typeform, you can show how many questions are left, which is huge in getting people to follow through. At times I was wondering how many more questions I'd have to answer.
In your results section, I would consider moving "why did beaker select this for me?" to the top. This is your USP and why the user is here. And you could show that right away if using a typeform. (also, seeing "would we buy this? "Meh" in giant letters isn't reassuring as a user).
I would also think about adding a subscribe button for people to be notified when you have new surveys / product types available.
-
2
This is great insight, thanks for taking the time to type it all up! Tbh the chatbot was very spur of the moment. At first the skincare survey was a simple form, but then I changed to a chatbot to give it a bit more personality and "interactiveness". I agree I may have made the language a bit too cutesy at the cost of clarity; working with a copy editor now to make the language a bit more universal.
To your very good point about not seeing the options that fit you: skincare quizzes are fluffed up with answers that you think have meaning. In reality, product chemistry is way simpler than that, with only a few variables. My first hypothesis was that people would be relieved to take a simpler, more straight-forward quiz. But I realize now that people are so used to the common choices provided by other brands that the simplicity confuses or frustrates them. This is another reason to change to a more typeform format: easier for people to type in/select what they want to tell me about.
This project has been a great lesson in how to innovate by only changing one variable at a time, otherwise your innovation gets lost in confusion. Also a great lesson in how much the format/method of information delivery is just as, if not more, important than the information itself.
-
-
1
Nice insight. Thanks for sharing.
A few questions:
Are they interested in the chemistry logic of how the recommendation was arrived at, or why that recommendation was given in layman's terms?
Are you dividing your customer feedback into groups in any way?
-
1
I've just gotten enough users to start dividing, yes. I'm now focusing on who's spending the most time on my site, what they're doing, and where (what referral channel) they came from. These are the users I want to best serve. I think this focusing should help me better edit/remake the recommendation system.
-
-
1
I think this makes a lot of sense, and the reflection you’ve done sounds really valuable and insightful.
I don’t have any advice to offer, but I will offer you some encouragement to keep going. It seems like it’s just a matter of time until you reach the point you’re looking for. Good luck and have fun doing it!
-
1
thanks!!! (even if frustrating building your own thing is SO much fun)
-
-
1
Good learnings. It’s easy to get trapped by what we think is best for others instead of leaning on what they actually say they want.
I think your spot on with your next steps. Building content to help folks better decipher all the crazy ingredients sounds awesome to me. Through the process hopefully the next product becomes more obvious and we’ll founded.
Super pumped to hear back about what you learn! Keep us updated!
-
1
Thanks! And will do!
-
-
3
This comment was deleted 7 months ago.
-
2
These are GREAT resources, thanks so much for sending! Paul Graham essays are great. I read that exact essay a while ago, and then again this past Monday. I think it put me in the mindset to finally start.
-
