#02 Slay the Assumption Dragon & Create a Research Plan
Founders are resistant to taking the time to do research– especially early stage when they have a product idea they are running toward or they have launched and just want to grow.
Over the years working with founders I’ve noticed there are a few key reasons–
It sounds expensive and time consuming.
There are only so many resources.
They don’t know where to start.
The most common reason is “I’ve already done research” but research is something to be done continuously. I’ll share more about continuous research in a future essay. This essay is about starting a research practice.
Research is often the difference between success and failure. And when you’re building a company or product with the intention of being community-first, it’s a non-negotiable.
Let’s go back to the objections: research is expensive, time-consuming and hard to plan.
I’ve developed a step by step guide to make a scrappy research plan. When I say scrappy I don’t mean not-valuable. I mean, smarter, not harder. Did you feel the objections poof into thin air?
Grab a notebook or start a new Miro board and let’s do this real quick. Why not now?
How to create a research plan in less than <30 mins
If you want to make community the heartbeat of your product strategy then the first step is to analyze what you’re currently building, planning and thinking.
Assumptions are the silent killer of communities and products.
Step 1: List out every assumption you're making in your business about your customers
Do this quickly and without thinking too deeply. You want these to pour out like journaling.
Here are some examples of assumptions:
we are solving the right problem
the onboarding flow gives users confidence in the product
the events hosted are at convenient times
XYZ feature is the best part of your product
demographic data of customers
Rule of thumb:
Is it a fact? No? Then it's an assumption.
Even if you're pretty damn sure. Write it down then go to step 2.
Step 2: Note your confidence level in each assumption
After you’ve finished your list, go back through your list and notate how confident you feel in each one.
Use this key to mark each one L/M/H
Low - I'm not very confident
Medium - I'm somewhat confident
High - I'm extremely confident
Check your ego at the door here. It’s okay if you’ve acted like you were extremely confident but now I have you second guessing yourself and maybe it’s more like a ‘medium’.
Step 3: Mark your assumptions as qualitative or quantitative in nature
Anything you marked with high confidence is not a priority for your research right now. Set those aside or cross them out and focus on what you marked as low or medium.
The next time you do this activity (I recommend at least quarterly) your L/M/H may shift and you can tackle those problems when you get there.
Separate the assumptions you marked low and medium by whether it is qualitative or quantitative data. This will help you determine the method you need to use to capture the data. For example, you can't get qualitative data from a survey. We’ll cover this more in step 4.
Qualitative assumption examples:
People would prefer to listen to their podcasts and music in the same audio player
Our customers are most overwhelmed with the operational aspects of businesses
People prefer a digital reading device while traveling instead of a paperback
Quantitative assumption examples:
Our customers are between the ages of 23-45
The majority of people in our niche are using apple devices
Our customers are mainly promoting their business on Facebook and Instagram
If you find yourself wanting to ask why or find the deeper meaning behind a metric, then move that quantitative assumption to qualitative. Sifting through qualitative data in a survey or analytics (1) isn’t effective and (2) is why we think of research as tedious and time consuming.
Step 4: Find or capture the data
Note each assumption with where you can find the answers to validate if it is accurate or not. You can get really creative here. Notice if your assumption is a broad data point about the world that you can find in other trusted research reports. Remember: smarter, not harder.
You might already have the data:
Stripe export (quant)
Analytics tools (quant)
Customer service emails (qual)
Community posts (qual)
Or plan to get it:
Survey
1:1 customer interviews
You’ll likely find yourself marking many of the qualitative questions as 1:1 customer interviews. These can be hard to recruit for and time consuming. If you can prioritize it, awesome. You’ll typically see trends at ~5 interviews per persona or customer type.
If this isn’t in the cards for you right now another great option is diving into community data. Regardless of if you’re doing 1:1 customer interviews, it is an excellent supplement to qualitative insights.
If you have a community or your customers hang out somewhere online– like a subreddit– then you may be able to find the qualitative answers you’re seeking there.
Try to find other posts and conversations that support (or don’t) your assumption. Avoid making a post with your question because as soon as you have one comment you’ll risk follow-the-leader syndrome or “group think”.
I find the best method is to take screenshots of the posts and pull them into a Miro board so that you can add notes on them and review for trends.
Step 5: Create a quick roadmap & delegate research tasks
Make a quick list of tasks from the last step. Keep your assumptions with each task so that you remember why it’s important and you or your team can tackle each assumption individually.
It will look something like this:
Export stripe dataset
Assumption 1
Assumption 2
Review Google Analytics for XYZ
Assumption 3
Review customer emails between <date> and <date> for themes regarding:
Assumption 4
Assumption 5
… you get the idea.
If you use a tool like Asana or Notion I’d build this list right in there, then it will be easy to assign tasks out to you or your team as soon as you’re happy with your list.
Don’t forget to schedule time on your calendar to analyze the results once you've captured all the data. Carve out at least 90 mins to review each data type.
If you do user interviews, those will probably require at least a few hours to analyze. If you can set aside an ‘analysis’ day I would highly recommend it.
My team has an analysis day once per week. It’s valuable because it turns into future content, “micro-pivots” as I call them and a more connected community. More on that in a future issue!
If you try this I’d love to hear how it goes for you and if you have any questions I’m happy to answer them! Send me a note or if you’re subscribed to my newsletter just hit reply 🙂
If you’re not subscribed to my newsletter yet, get the latest essay + other resources every Friday.