Startup Superpower: Customer Discovery Machine

I’ve worked in software product companies for over a decade in engineering, product management, and product marketing. Having led both, product management and product marketing at different organizations where i dealt with customers day in and day out, I had still not learnt proper customer discovery. I realized the importance of discovery process once i began working on my startup. In fact it is so basic to a startups that without this skill it is hard to succeed.

At Berkeley Skydeck, this is one of the most valuable things I learnt through an NSF iCorps course on this, which is very well run.

Once you have a well oiled discovery machine, it’s like having a superpower that you can turn on anytime you want, and get market data flowing into your startup.

Note: This is in the context of B2B software company to Fortune 500

What is Customer Discovery:

As a founder, you want to figure out what are the top 3 problems of your customers, for which they will pay big bucks. The key to finding out is simple - during the conversation, don’t talk about your product, don’t nudge them, listen to the problems they volunteer. If your solution does not solve a problem that is volunteered by the customer in the top 3 priorities, you’ll find it super hard to sell.

Here is the crux of it — to get your product into an enterprise requires jumping through numerous hoops, and you’ll need an enthusiastic customer-champion for your product who will push it through their company. Some of the hoops are endorsing your product in front of management, getting the budget in her Enterprise, getting approvals from sourcing, project management and security. No one’s going to push for their number 7 priority, or nice to have.

In my case, and possibly in yours, the customer is not simple, the users are data engineers and the buyer is an executive. Now, what are the top pain points of each stakeholder? Next, you need to understand the top pains of your executive buyer, the person writing the check, and for which they preferably have budget allocated. What is their biggest priority? Many of these come through themes such as cloud migration or digitization (yes, this exists).

Customer Discovery Process

Your goal is reaching a steady state of 5+ conversations a week, at will. In the beginning, one needs to start at 10 conversations a week, this will ensure a good process. Try to get as many as possible in person. I’ve ended up with a majority on phone or video call.

How do you get the conversations setup? First, reach out to people you know and your friends know, these tend to be easier to get over coffee.

Next, use LinkedIn to search for people who match your target role and industry, and send them an invite with a short message, two lines and simple. It should be clear that you’re trying to understand challenges in the space and not going to sell a product. Many people will want to connect with you since you’re a founder, doing exciting stuff, and you politely asked for their help. I usually get a 20% initial response rate and that is the higher end — if you get this you’re off to a good start.

Once connected, send a second message asking for a time to chat. Ask for 15–20 minutes, it’s hard to say no to that! And they will always end up talking for 30 minutes at least, especially if you’re listening and they’re talking :) For someone who is a few years out of school, you can send a calendly link and they’ll pick a slot. People more senior tend not to like this, ask them for a couple of slots and take one. Here you should be hitting 7–9% response rate (of the original invites)

On the call, introduce yourself, just the broad area you are working on, do not introduce your product. Ask them about their problems, what does their environment look like, who does what in their team and company, what tools do they currently use, what’s not working well, and what could you solve for which their organization will pay big bucks. Take notes about the data, not your opinion of the data.

Hypothesis Refinement

You’ll start your discovery with clearly written hypotheses, that can be clearly negated. Now you’ll find out only some of them are true. So you’ll update your hypothesis for subsequent calls, this is an iterative process.

The purpose of first discovery calls is to get a good sense of the market and the problems that are important to many customers. I felt that after 30 calls, I had started to learn lesser on the new calls. If there is a lot of refinement, it’ll take more.

When to run the discovery process

The simple answer is always be discovering, there should be one founder who is always intimately connected to the market beyond your existing customers.

In the first pass you’re trying to discover what product to build. Once you get first design partners (as you might call your first few customers), they’ll give you really good data for their use case, now that they have skin in the game. But everytime you come to a bigger decision point on a feature, you want to know whether you’re becoming a development shop for this customer or are you solving a problem many people will want solved.

You’ll use targeted discovery to figure out what features are important to many customers. If there are incumbent products you’ll find out what customer think of their implementation of those features.

Hope this helps you set up a good discovery machine, a superpower that you can turn on at will, to keep a pulse on the market and help major decision points!!