Moving into week 4, the teams are now one third of the way through their time to build. Week ten, July 23rd, is demo day! To match the urgency and the nature of the accelerator, I’ve asked that each team commit to completing two experiments this week.
Only run experiments that can be completed in one week
After we went through the normal news items (feedback on mentors, books to read, new places to get dinner, etc), each team presented the results of their experiments. A recurring issue we had this week was that the teams bit off more than they could chew, such that no clear results or learnings could be had from their experiments. The solution to this problem is to only schedule experiments that you can actually achieve in one week (the cadence I chose for the accelerator)!
Here are some examples. One team wanted to approach a possible funding source, as a test to see if their idea was advanced enough to receive an investment to spur development of functional prototypes. They set up their test to pass if the funder, Elaine Hamm at i2e, would agree and commit to presenting their business to the board. This is a great test, because it shows both a serious commitment from a potential funder and is something that can be completed in a week.
Another team wanted real estate agents to invite them out to have their properties video taped by drones, and to have the video posted to the real estate agents website. However, the team only got verbal commitments from the real estate agents and the agents didn’t get around to inviting them out. Technically, this test failed. But the team is hoping that the agents will invite them out the following week, so we didn’t actually learn anything. A better test here would have been to get commitments from agents as step 1, validating that the agents see a need for this service. Then a follow up test could be how many of the agents that commit invited the team out during the following week.
What’s the point of testing again?
Now that the teams are in the thick of validating their ideas, I feel like we are starting to slip towards running tests for the sake of running tests, just to see what happens. The purpose of an experiment is to validate (or invalidate) a hypothesis. And the hypothesis should be tied to an assumption on the business model canvas, hopefully the most critical assumption.
Trough of despair
Even as the teams start to pick up the pace and sell their products to get feedback from their customers, Steve Blank warns of the trough of despair that comes from the need to pivot based on negative or lukewarm feedback. Teams may find that their original idea wasn’t as good as they had originally thought. My opinion is this more a cause for celebration! Rather than wasting another one to two years of your life and who knows how many resources, teams can commit to shutting down one option and pursuing another that may turn out to be more lucrative.
This is a revisit of the search vs. execution concept. The more side paths you can visit, or the more swings of the bat you can take, the more likely you are to find the one that works. While in search modes, startups (or young companies), need to be spending their efforts on finding product/market fit. Once that happens, then they can start switching to execution mode.
Discussion of channels
The heart of the curriculum today was distribution channels. That is, the way customers are going to get their products or services delivered. We had a brief rundown of the types of channels available, primarily direct, indirect, and OEM. I encourage startups looking for product market fit to start with direct channels because it comes with the added benefit of getting direct customer feedback. Ultimately it is important for the startup to know how their customers will get their products. While there may eventually be many ways (you can rent a movie from netflix, stream it on Amazon, or go to Redbox to pick it up) or channels to choose from, a startup serves its interests well to validate one channel at a time.
Many teams avoid indirect because they feel like they are giving money away that belongs to them. However, indirect can dramatically reduce selling costs. For example, the wheelchair team is wondering if it would be better to sell their device directly to consumer or use well-known wheelchair shops. While the upsides of direct are straightforward, building out your own sales channel and making sure you get enough traffic is a substantial cost in and of itself.
Distribution channel economics
The first step in building a startup is to get traction. That is, where people are interested in what you are creating and will make a commitment to your young company (the crazy ones!) A step you need to take shortly thereafter is to figure out if you can make money selling your product or service. And a big question for channels is cost per acquisition, or how much will it cost to get a new customer in the channel you’ve selected.
This image is how I drew out a typical sales cycle. There are all sorts of charts that I can’t ever remember available on the web (like this sales cycle chart).
I prefer to use the marketing funnel as you can see from the whiteboard illustration, that demonstrates how a person becomes a lead who becomes a client. Layering on top of that is Steve Blank’s idea to demonstrate understanding through drawing. Here’s what the parking team came up with.
In short, the sales cycle describes everything that needs to happen to find leads and convert them into customers. It tends to fall into categories like initial contact, first meeting, followup and sale. After a lead becomes a customer the sales cycle may continue with referrals and upsells and future purchases, but those items are not crucial to figure out at this stage of the game.
Some key questions to consider when thinking about a sales cycle or marketing funnel are:
- How long will it take to get a new lead to become a customer?
- What’s the lifetime value of a customer? (impacts how much you can pay to acquire them)
- How do you prove a channel? (prototype it! get real intent to purchase)
After building out our sales cycles and drawing them, each team decided on which distribution channel they would focus on first.
Creating new experiments
Class closed as usual with the teams working on their new experiments for the upcoming week. For week 4 they would again need to have two experiments to work on. Going into week 5 I’m going to challenge them to up their velocity and try to get three experiments in per week for the second half of the accelerator.