How we invalidated an idea and completely pivoted our start-up

This story started in September 2017 when I came up with an idea of a chatbot to replace an app which was developed a few years earlier for one of my customers.

The app provides prepaid coupons for drinks in a network of various cafés across Saint Petersburg. Folks who ordered the app were trying hard to get their first digital business running, however, the coupons sales were so-so. So I had an idea that bringing up a new trendy technology would change the situation and boost the business.

With this idea, I and my friend and colleague Robert applied for the Elevator Startups program which was about to start on 5th of December. Soon we were accepted and the journey has begun!

Our new mentor from the elevator, John Sechrest, pointed out that the chatbot idea needs to be carefully validated before going to the implementation. John recommended reading a book called Lean Customer Development.

After taking a number of interviews with business owners in December I was convinced that the problem of the coffee folks is not in technology.

The interviews which I took were running under the Adam Berk’s @adamberk experimental incubator that gives startups and entrepreneurs funding in exchange for testing and sharing lean-startup hypothesis. The individuals who helped me run this experiment are John Sechrest, adamberk and Robert Filić. The link to my specific hypothesis on trello is THIS.

In order to formalise the results and complete the experiment, I will now follow the structure given by Adam.

EXPERIMENT

I THOUGHT that 6 out of 10 cafe owners who do digital marketing and spend at least $100 per month, will subscribe for early access to an alternative digital channel by the end of the next week (22.12.2017) because it’s too complex for cafe owners to define parameters of their target audience on social media platforms they’re currently using.

ACTUALLY, all the interviewees committed using an alternative solution. In total, I have had conversations with only 4 people who match the segment targeted in this experiment.

Findings and Conclusions

Although the results may look quite optimistic, we came to a conclusion that the idea isn’t worthy of implementing and INVALIDATED the hypothesis. Below are some findings from the interviews.

— Cafe’s location is probably the most important factor which impacts on the number of visitors. A rare person will travel half an hour to visit a cafe even if it’s superb.

— There are two types of cafes differentiated by their marketing strategy:
a) ones who almost don’t spend anything and does only SMM (groups). In order to compete with other cafes, they rely on word of mouth and try to beat by quality/ price/location;
b) ones who have a dedicated marketing person and spend a good budget on it. However, they say they ARE NOT struggling with their target audience definition and with ad postings. These cafes often run advanced loyalty programs and buy an expensive license for a loyalty service.

— For all the cafes the existing digital solution from our partners doesn’t bring any extra value — just a few occasional visitors. Since the service doesn’t cost anything, they can keep it going. Those who have some traffic, aren’t quite happy too since they say they need new customers from the app but not the returning ones.

Finally, we quickly realised that the partners aren’t doing a good customer development work and don’t keep in touch with the businesses on a regular base. Every taken conversation brought more confidence that changing a digital channel (from an app to a chatbot) will not bring any significant value to the coffee businesses.

What’s next

We changed the domain and made a new hypothesis. In a next experiment we are going to prove that chatbot technology can bring value in fashion business by providing personalised style recommendations and helping people to find and buy clothes online.

Since we already have some number of people who did a sing up, we’d run an experiment with stylists and spend $500 on purchasing clothes and re-selling them to the users who subscribed for an EAP. Basically, that kind of an MVP should confirm the viability of the new “Personal Shopping Assistant” idea.

The NEXT EXPERIMENT is submitted under the following Trello card.