Checklist for a Thorough In-Depth Study on a Discovery Stage

Anastasia Berezhnova, Head of CustDev Center, explains how to conduct a deep research of consumers’ pain points on a product discovery stage. Read this article for an experience-based advice on formulating hypotheses, analyzing the market, generating leads for interviews, launching a lead funnel, and more

On a discovery stage, any startup team goes through a process of seeking for viable product hypotheses. Since there is yet no product to develop and sell, startuppers need to find such acute consumers’ problems that people will want to acquire the solution to them in exchange for their hard-earned money.

In this article, Anastasia Berezhnova shares her checklist that contains all the necessary steps of an in-depth study on a product discovery stage. It includes not only research and leadgen recommendations, but also practical advice on preparing for and analyzing CustDev, JTBD, and hybrid interviews.

Step 1. Formulate a hypothesis

There are different types of hypotheses — for example, they can be related to market analysis, lead search, or how factor A affects metrics B, C, and D. This article primarily talks about hypotheses that one formulates when building a startup.

Most often, product managers come to their teams with extremely generic requests, something like, “I want to make a furniture rental service in Europe.” But this is not a hypothesis. A well-formulated hypothesis should always contain an If-Then statement. Here’s a universal formula for it:

If we do X, then we’ll get the result Y.

Possible results may include

• various benefits for a startup (for example, 200 new sales),
• benefits for customers (30% increase of their business KPIs),
• new leads (+1000 new customers, +500 newsletter subscribers),
• confirmation of consumers’ pain points, and so on.

In the experience of Admitad Startups, we formulate several hypotheses at once and proceed with exploring them during in-depth interviews. For example, we might be trying to validate our assumptions on these topics:

• What problems do consumers face every day?
• How do they cover their needs?
• What tools are used for this purpose?
• What do consumers find lacking in these tools?

Based on these deep interviews, we can create truly appealing offers that would bring us customers and, eventually, money. The more hypotheses we validate, the more we know about our target audience and the more likely we are to build such a product that consumers will agree to buy.

Step 2. Set quantitative goals and criteria for leads

After the hypothesis, it’s time to set quantitative goals and determine the lead criteria. 

• Lead criteria are assumptions about segments of the target audience that the team makes before diving into deep research. 
• Quantitative goals are required in order to understand how many people will be interviewed from each segment.

It’s better when the lead criteria are flexible, whereas quantitative goals are fixed. Admitad recommends conducting 10–15 interviews. 

Why is it important? A product manager can request 10 in-depth interviews, but the 10th interview happens to contain some unique idea that changes the PM’s hypothesis. So they request 10 more interviews, and it goes on and on and on. So in order to avoid these infinite cycles, you need to set specific goals. When you hit them, you’ll know that it’s time to stop.

As for lead criteria, in-depth interviews may or may not refute your assumptions about the target audience completely. It’s alright, nobody is trying to guess all the segments. Instead, these criteria need to become a starting point that your team uses to plan the outline of the study.

Also, since a startup usually has limited resources, its team can’t survey everyone. So before a lead is invited to an in-depth interview, they must answer a bunch of qualifying questions. It’s a short screening that determines whether this lead fits the criteria and qualifies to become a potential customer. This way, a startup team ensures that they only talk to relevant leads.

Example: furniture rental

Here is a live example from the experience of Admitad Startups. We were building a furniture rental service for the US market called Rentomatic. American citizens mostly rent empty apartments, so furniture rental services are quite popular there.

When we were looking for lead criteria, we came up with several possible segments. One of them was students, ages 18–30, from certain states. We figured that they needed rental services because they

• went to university/college across state borders, or
• had no money to buy furniture, or
• shared apartments with friends.

So we started looking for students at the right age in the right states. When we collected a list of potential leads, we proceeded to screen them with qualifying questions, trying to understand whether they were relevant for our product. 

There were three key points that we asked:

• How long ago was your last move?
• Have you ever tried to rent furniture?
• How much did you pay for the rented furniture?

If the student moved many years ago, they were irrelevant for our research since they no longer remembered their customer journey. We also crossed out those who paid insignificant amounts for furniture rental. 

Step 3. Analyze the market

When a startup team is first given an assignment, most of the time they are not familiar with the topic. For example, our colleagues that started working on Rentomatic did not know a thing about the furniture rental market in the United States. They had no idea it was a common thing for locals to rent empty apartments and use furniture rental services.

So when you take on a new task, you need to start with the market analysis. Be ready to set aside 1–2 weeks in your work timeline to study the competition, gain insight into the niche, and write research materials. This should be your ground zero.

It would make sense to search for analytical articles and statistics that describe market conditions. Another easy method is text surveys and questionnaires. There are various respondent platforms (for example, Respondent.io) that allow you to post questions for the general public. As a rule, they also allow you to select the segments of your target audience that you want to reach. 

Overall, market analysis is exactly the stage when a startup team gets roughly acquainted with the industry. Further steps become clear in the process. For example, when we researched the market before launching Rentomatic, we posted this list of questions:

• Have you ever rented furniture?
• Why did you rent furniture?
• What pieces of furniture did you rent?
• What website did you use?
• How many times did you move?
• How many of your friends have moved?

We also found articles stating that the average American family moves 11 times in their lifetime, and that 85% of rented apartments are unfurnished.

Step 4. Find ways to generate leads

As soon as the market analysis is ready, a startup team can move on to the next step. Now it’s time to deduce several different hypotheses about where a startup will find leads. Aside from paid platforms that connect businesses with target respondents, there are always less obvious options, and it’s best if you can cover them all.

For example, Rentomatic was testing these assumptions:

• Customers of furniture rental services leave their feedback, so one can learn their names from reviews and parse their profiles in social media.
• There are social media chats and groups of people who are looking for furniture rental.
• People who are looking for an apartment are also looking for furniture, so one can partner up with apartment rental services.

Every lead generation hypothesis should be quantitative. For example, “If we post 5 ads in chats for people who are looking for furniture, then we will get at least 5 interview appointments in 2 weeks.” It’s important to limit the time frame so as not to get stuck.

To keep track of everything that is going on on this stage, you can make a table with columns: ‘Actions taken’ and ‘Results obtained’. This is an easy way to see which hypotheses brought the most leads and which ones should be discarded as ineffective.

Step 5. Prepare scripts in advance

For lead generation in each segment, generic text messages get prepared beforehand. These are scripts that follow scenarios of each possible conversation with respondents. Thanks to them, startup employees will always know what to say next in order to get the maximum benefit from communication with leads.

Scripts must be translated into all languages necessary to conduct in-depth research in a particular market. For each segment, a separate text message is written in advance. Here are some categories that Rentomatic singled out:

• Message for social media chats
• Email for partnering rental services
• Private message to cold customers
• Private message to warm leads, etc.

To write good scripts, you need to follow a customer’s journey from A to Z. Ask questions about every turning point, starting with a moment when a customer decided they needed a product, and ending with reasons why they chose to purchase from a particular store. 

Step 6. Launch your lead funnel

When all of the above steps are completed, you can launch the funnel — i.e. all the channels of lead generation in order of priority. It’s best to start with the ones that perform the best across your hypotheses, ending with the ones you believe in the least.

At Admitad Startups, we prefer to launch lead generation channels in the following order:

1. Respondent.io (it takes a long time to set it up)
2. Social media chats
3. Manual lead search
4. Paid lead parsing (if purchased)

Step 7. Conduct your in-depth interviews

When your funnel starts bringing in new leads, you begin interviewing them. With cold leads, you usually need to ask some of the qualifying questions. Warm leads mostly come from paid platforms, so they have already been screened.

Be cautious of paid respondents. Sometimes they pretend to be whoever you want them to be because they earn money for each interview they participate in. So these leads need to be checked for authenticity — find them in Linkedin and/or Facebook beforehand and make sure they are a real person.

Of course, if you realize that the respondent is lying to you during the interview, you can contact the platform and get a refund. Still, it takes little detective work to filter out respondents who seem ingenuine — and it will save you even more resources in the long run.

Now, let’s talk about interviews. One meeting usually lasts between an hour and an hour and a half. If you can’t be present in the same room, think of conducting a video interview in Zoom, Discord, Google Meet, etc. — this way, it’s much easier to read a person’s emotions. The entire conversation should be recorded and stored so that you can return to it later.

Step 8. Transcribe the interview material

Transcripts are required to complete the Jobs to Be Done research. Instead of only taking up this task when you finish conducting all the interviews, you can start transcribing them in parallel. For example, hold one or two conversations and immediately transcribe them. This way, you can swiftly notice the interviewer’s errors and edit (if needed) the list of questions.

It’s useful when the person who transcribes interviews is not the same person who conducts them. An outside observer can point out unobvious gaps, poor wording, or questions where an interviewer needs to dig deeper. 

At Admitad Startups, we have two methods of working with texts: transcribing them and writing a detailed summary.

• Transcripts — for JTBD-only interviews. This is a machine-performed, automated task that needs to be rechecked by an employee.
• Detailed summaries — for CustDev interviews and JTBD+CustDev hybrid methodology. These are done by a live person who analyzes the information and provides their feedback.

Step 9. Provide the interim report

This paragraph describes how we do it in Admitad Startups. Your processes might be organized differently, but still, an interim report is a useful tool to track your project’s journey.

In the middle of our research, we usually meet with the product manager and talk over the results that we’ve got. There are 3 possible variations of how this discussion might progress.

1. The original hypothesis has been confirmed. Product manager approves further work.
2. The original hypothesis has been refuted. In other words, we suddenly realize that the problem was somewhere else. To save time and budget, we change the hypothesis and reshape our future product so that it conforms to the new conclusions.
3. The original hypothesis has been refuted. However, the product manager insists on keeping the old one until we confirm that we’re going in the wrong direction. If our failure gets 100% confirmed, we change the vector of our actions.

Step 10. Collect and analyze the results

When the transcripts are completed, we distribute the data between the following documents:

• Map of pain points
• Jobs to be Done canvas (more on that in our article)
• List of customers/leads

The last document is a table where we sum up the results: the number of our customers, their lead criteria, their income rates, tools they use, money spent on them, etc. In short, the list contains everything that’s important for us to know about the lead segment of our research.

Let’s illustrate it with an example. If your lead segment consists of bloggers, then you might want to include

• which websites these bloggers come from,
• how their subscriber base is segmented, 
• how they monetize their subscribers, and so on.

Final thoughts

Anastasia Berezhnova shared her checklist of a deep research of consumers’ pain points that is usually conducted on a product discovery stage. Her advice comes from the experience of Admitad Startups and covers the way we conduct in-depth studies in our studio.

Here is a short summary of all 10 steps that Anastasia suggested.

1. Formulate a product hypothesis using an If-Then statement that mirrors the desired results.
2. Determine your lead criteria, i.e. must-have attributes of your respondents, and the number of leads you will interview.
3. Get yourself familiar with the market, niche, and competitors before you dive into deep research.
4. Formulate hypotheses on how you are going to generate leads and discover your leadgen channels.
5. Prepare scripts to engage with your leads. Write down all possible communication scenarios for all segments of respondents.
6. Launch your lead generation funnel, adding the leadgen channels in order of priority (from best-performing to least-performing).
7. Conduct in-depth interviews: CustDev, Jobs to Be Done, and/or hybrid.
8. Transcribe the interview material into text format so that you can see and highlight the most important points. 
9. Report intermediate results of your in-depth study to your product manager to determine whether your findings align with the original hypothesis, and what you should do in case they don’t.
10. Collect the results of your in-depth study and analyze them: map out consumers’ pain points, fill in JTBD canvas, collect your lead list.