This is the second part of a two part series on how to do basic customer research. The first part covers preparing and running research, please check that out before this one (or it won’t make a ton of sense)
I know, I know, I promised you that you’d have this edition last week. Not gonna lie, I’ve been struggling a little bit over the last few weeks. I’m a smidge overwhelmed by life goings on. So this is obviously a little late.
The steps
Now you’ve completed your research, you’re going to want to analyse what you’ve learned and put it to good use. Much like a tomato, you can create lots of different things with your research, but the starting point is always pretty much the same. I’m going to walk you through how to analyse and some different ways to use what you’ve found out.
This week’s lesson, looks a little like this:
Purpose - what do you want to do with what you’ve learned?
Analysis - what is it, why do you need to do it?
Manual analysis - the osmosis approach
Analysis using AI - The fastest method
I’m not going to go into the detail of creating the individual assets today, because you’ll see from the list below, it’s LONG… and I’m writing this in 27 degree heat when I promised myself a day off reading in the sun.
1. Purpose
How you want to use your research lines up with your learning objective from the research you’ve just done. It’s the yin to the yang of your learning objective from the first part of this series. You’ve been clear on what you want to learn, now you just want to be super clear on how you want to use it.
There are, however, a lots of ways you can use your research. Nailing this now means you won’t miss anything in your analysis stage.
Here are some examples of how research can be used or contributed to create important resources that power your business. I’m calling these assets (helpful as you read through the rest of this week’s part).
Profiles
Customer
User
Strategy
Sales
Marketing
Product
Features
Content
Design
Product i.e. what features or functionality you want to include
Service i.e. your pre-sales or customer support services
Content
Tone of voice
Topics
Business Model Canvas
Value proposition
Customer relationships
Customer segments
Channels
Note: you might be wondering what the difference between a customer and a user is. A customer is someone who pays you money for a product or a service. A User is someone who uses something you’ve created, the two aren’t always the same. Using the term customer puts a sales and marketing focus on your research. Using the term user puts a features, functionality and usability focus on your research. The distinction is important when you’re analysing your findings
Don’t worry if you’ve done the research and don’t feel like you’ve got enough to answer all of the above. These are just some examples of how you could use them. I will probably do additional sessions to cover how you can use research to generate the topics listed above.
Using the two examples I gave you in the first part of this series, let’s look at how you would use research to meet those needs.
Starting a business
If you’re starting a business, your research is going to give you a massive leg up to help you go in the right direction. The ways that you should use the research in this case are:
Customer profile
Strategy
Sales
Marketing
Product or service (if you’re a service based business)
Content
Product or service design
Content topics
Business Model Canvas
Value proposition
Customer relationships
Customer segments
Channels
If this list seems seriously overwhelming. Start with the customer profile.
Launching a new product
If you’re launching a new product, or even updating an existing one to better reflect your customer base, then these are the ways you should use your research:
User profile - if the user of your new product differs from your old one, this is a good starting point for you.
Strategy
Sales
Marketing
Product
Product design
Business Model Canvas
Value proposition
Customer relationships
Customer segments
Channels
I’m going to assume you’ve already got a customer profile. But as ever, I might be making an ass out of myself assuming this, so if you haven’t, then please start with that.
Why this step is important
It might seem a little weird writing down how you plan on using the research you’re about to analyse. Getting clear on how you want to use it means that when you dive into the research you’re looking for things that help you to shape the thing you want to create.
If you haven’t made it clear you want to know how your ideal customer describes something, when you come to write content for it you’ll either have to do this again (totally ball ache) or you’ll have forgotten you’ve got this detail you can tap into (seriously annoying).
Good prep at the start of creating anything means you’ll have a far more successful outcome. It’s the same thing as cooking a super complicated recipe without having prepped all the ingredients and read the instructions several times. You’ll miss an important ingredient or burn something. Best to avoid that, yeah?
2.Analysis
Now you’ve got super clear on how you want to use your research. Here comes the fun bit. Analysing what you’ve learned.
Analysis can be time consuming, as you need to go back through all the research you’ve done to extract the insights from it. I’m going to show you how to do it manually and how to use AI to help you do it.
When it comes to analysis, the ratio of time to do the work looks something like this:
- Manual analysis 2:1 - two hours analysis (minimum) for every one hour of research you carried out
- AI analysis 1:1 - one hours processing for every hour of research you carried out.
What you’re doing with each type is slightly different. With Manual analysis you need to extract the insights and then analyse what you’ve got to create an aggregated view that can be used to create individual assets listed above.
Thanks for reading this edition, if you’re not already a subscriber consider becoming a free or paid (and making this author’s day) member.
2a. Manual Analysis
OK, I’m not going to lie, this is time consuming and will require you to switch to analysis and questioning mindset. There is nothing wrong with this. You will absorb a lot more of the information, and will be living and breathing analysis tool.
Materials
- You’re going to need some post-it notes and pens,
- or a virtual version of them (i.e. Mural, Miro, LucidChart or another tool).
- You’re also going to need your notes (if you took them), re-watch your video (if you recorded it) or read your transcripts (if you have them).
Not gonna lie, if you haven’t got any of these, you’re kinda fucked. It’s why I suggested recording the session in the last part.
Observations and insights
Using your notes, video or transcripts, read through each one and use one post-it note for each thing you observe.
Here are some of the things that I look for when I’m doing this approach:
- Frustrations - things they say that annoy them or they absolutely hate about something
- Fears - what are they afraid of happening or doing
- Passions - what things do they love about a product, service, way they do something, things they do
- Wishes - what to they wish they could do, or for a product to do
- Jobs - what jobs are they doing *note: a job never changes, how they do it or the tools they use to do it can and will - it’s also where your opportunities lie*
- Quotes - things that they said that *really* captures something above, grab those they’re like gold dust and much harder to capture when you’re using AI to analyse your research.
- Tools - what tools are they using or wish they were using
- How they do things - document the step-by-step of how they do something
- Problems - what problems are they experiencing in the way they do something?
Essentially, the things you’re looking for are signals you’ll use to help you get answers that allow you to shape the direction you want to go in, and the assets you want to create from the end of it.
If you’re using a whiteboard tool like Mural or Miro to help you do this, you can use labels to categorise your post it notes. If you want to use the list above as your labels apply them as you create them. This will help you with the next stage.
But before you go to the next stage, stop. I mean that in the best possible way. You’ve churned through a lot of information, I want you to step away from the analysis for at least 24 hours. Your subconscious is going to do some heavy duty osmosis work, if you let that bad boy loose, the next step will be a lot easier.
Sorting
Once you’ve gone through each transcript and created your insights, and hopefully categorised them with labels along the way, your next step is to sort them.
The purpose of sorting them is to look for patterns in the information. Patterns give you laser point accuracy of where to focus your efforts.
If you’re using a whiteboard tool you can get it to sort them by your labels for you.
Once sorted, start looking for where you’ve heard people say the same thing. These are patterns (obviously). The duplicate nature of them tells you things like:
- Where to focus
- Really big problems your customers are having
- Feature priorities
- Language to use (i.e don’t call something pink when your audience uses flamingo pink)
Sorting helps you to reduce what you’ve found into invaluable and valuable insights. Those are the ones you want to focus on when creating your assets.
Other things to group
Thinking about how you want to use your research, you now might want to start grouping them into areas that reflect the assets you want to create.
I’m not getting into detail here on how to do that (see earlier comment). But, if you were planning on using this research to help you create a business model canvas, use the titles of each section of that canvas to help you sort through the information.
Thank you for reading Unperfect Choices. This post is public so feel free to share it with anyone else who you think this could help.
2b. Analysing using AI
Before you get your knickers in an absolute twist at me suggesting AI, if you’ve never analysed research before, time is not on your side or your brain just isn’t wired in an analysis way, AI is an absolute saviour.
I’m cheating slightly here, because above I said I wasn’t going to tell you how to create the assets listed above. If you use AI, you can skip the analysis and go straight to creating the assets. Which is fine, but I would strongly implore you to ask it to do the insight extracting listed above first, before you ask it to jump in and create the assets. You might be asking me why I’m trying to punish you like this, but it’s for a good reason. Understanding all the emotions listed above before you create an asset with them helps you to walk in the shoes of the people you want to work with. That’s the whole point of research, so taking a hot moment to bathe in all that knowledge will do you a solid favour in the long run.
AI is both good and bad. I have mixed emotions on it. However, it is the cheapest way for you to move forward in your business when you don’t have all the skills needed. It doesn’t replace paying an expert to help you, as you’ll get so much more from that, but only ego maniacs like Musk can afford to pay anyone and everyone (yet still ignore the advice). If you have a budget, pick carefully where to use it, and then backfill with AI to give you an extra pair of hands.
I’m going to tell you how to use ChatGPT and Notion AI to help you as they’re probably the tools you’re either already using or have access to.
ChatGPT
Important note: if you have carried out research that is commercially sensitive, contains information that’s under an NDA or is protected by privacy laws, DO NOT under any circumstances use ChatGPT to analyse your research
You will need a paid subscription to upload files and analyse information in ChatGPT, but it’s less than £20 and you can cancel after you’ve done this work.
There are two different ways you could go with this approach. You can analyse the transcripts individually or you can aggregate them and analyse together. There is no wrong way to do this, but I’d recommend doing both so you’ve got some good reference material.
Individual analysis
You can upload each transcript individually and explain to ChatGPT what the transcript is, what your learning objective was, and ask it to a. Summarise the research with this participant and extract key insights that relate to your learning objective and anything else it feels is relevant.
Here is what that prompt might look like:
Please analyse the attached transcript from customer research I carried out to learn about the problems my ideal client has that relate to my business which is (include a description of what your business does here). I would like you to summarise the research and extract key insights that relate to my learning objective and any stand-out findings, including how I might use them.
Aggregated analysis
In this step you upload all the transcripts, explain what they are (i.e. research transcripts), what your learning objective was and ask it to create a specific asset.
Here are some prompts you could use
Customer profile, jobs to be done and problem statement
The attached transcripts are from customer research I carried out to learn about my ideal customer and what problems they have, please can you analyse the transcripts and create a customer profile that highlights the jobs they do and the problems they are, I would also like you to create me a problem statement.
Business model canvas generation
The attached transcripts are from customer research I carried out to learn more about my customer and whether the product I’ve designed will help them solve their problems. Please could you create a business model canvas, including detailed value proposition, based on the analysis of the transcripts.
When you’re writing prompts, be very specific and clear in what you want it to do. Leave nothing to ambiguity.
The great thing about this is that it does literally all the hard work for you. You can also ask it to tell you how you can use the things you’ve found out.
Note: If you use Butter as your video conferencing tool, they now have ChatGPT built into their transcription tool. Which means you should be able to do all of these things without needing to download the transcripts or have a ChatGPT subscription. Very handy.
Notion
If you have a Notion AI subscription you can use it to do something similar. Notion makes it clear that your data is not used to train its AI. However, it’s still using a GPT model so if you’re super concerned you might not want to use AI to analyse your research.
Instead of uploading a doc with the transcript of your research, create individual pages for your research and paste your transcripts into them.
In the bottom right corner of the page you’ve created look for this icon:
You’ll get a pop-up box that lets you have a Q&A with the information on the page. You will need to pick the page you want to have a conversation with, so look for the page that you’re on, and pick that one.
Once you’ve selected the right page, ask it to summarise the information or create the assets suggested above. It will do that for you in the dialogue box, which you can then copy and paste into that page, or another of your choosing.
Unfortunately, you can’t do the aggregate analysis of each of the pages unless you copy all of the transcript into a single page. If you do that, you can then use it to create strategies and business model canvas details.
Other options
There are lots of other platforms out there that can do transcription (Otter is one) or AI analysis (like Microsoft CoPilot, although they’re going to stop offering it as a commercial product).
Whatever you choose, using AI will make this job a lot easier for you, and help you to see things that you probably won’t be looking for. It’s excellent as a spare pair of eyes.
Analysing research is time consuming, and can be boring AF. But it’s also the one that opens your eyes to all the possibilities of creating something that customers will throw money at you to get. Do we genuinely think that anyone at Tesla did research when they created a $700 sledge hammer? (It’s real, I shit you not). But equally, do we think Apple listened when they got a right bollocking about battery draining? (They did, even if they also got fined for it).
Designing products that people want to use and pay money for is more guaranteed if you do this bit of work.
I’m going to switch to creating these type of newsletters every two weeks. 3000 words takes about 3 hours to write and an awful lot of brain power.
I’ve also thought a lot more about the other side of what I want to write and who for. I’m still working out the details. But I know that it’s going to be more focused on the way I’ve actually been making money this year, which has to do with digital and business transformation. Don’t panic, it will be with my usual slice of “holy fuck I didn’t see that coming”. But it will be more transformation focused.
With that, I’m off to make a sandwich, ask the cat to move so I can lie down and enjoy the rest of the afternoon in the sunshine.
Emma
Related posts
Beginners guide to doing research
Emma Chittenden • Jun 13, 2024
Hola 👋🏻 As I promised (threatened?) last week, I’m going to dial up the helpfulness of the content I share with you. This week I’m giving you everything you need to do your first bit of research. This is my tried and tested approach to doing research. You’re getting 26 years insight on how I’ve done research myself, commissioned it in design teams, or…
Read full story →
