Establishing a robust product discovery infrastructure at your organization gives you the freedom to focus more of your time on creating, analyzing, and testing customer insights.
It also doesn’t need to be expensive, or fancy, to be effective.
Here’s how an infrastructure can help you
Let’s take an average example of a discovery initiative involving 15 customer interviews across multiple customer cohorts. If starting from scratch and doing things manually, you would need to:
- Categorize the different cohorts within your customer base
- Identify customers from those cohorts
- Recruit those customers to participate in interviews
- Request and process incentives for interviewees
- Schedule interviews
- Reschedule interviews – because everybody’s human and this always happens
- Schedule colleagues to do the interviewing
- Finalize objectives and guidelines
But wait – you don’t have your customer data structured in a way to easily filter by cohorts or a list of customers that have already indicated an interest in being interviewed?
And you’re telling me you don’t have guidelines for how to conduct interviews, analysis exercises, and communicate findings?
Doing all of these tasks manually drastically increases the complexity and time spent on this type of common research method. A lean product discovery framework can help make it all of this so much easier.
To help you see the benefits of having an infrastructure in place, this post is broken down into the three key pillars of the framework.
Automate processes to do the boring work for you
You’ll save a lot of time down the road by automating a few different processes or at least conducting them manually but routinely. Also, automating the boring work just makes everybody’s lives a little bit better, right?
Here’s the different ways you can set things up for success.
- Create a process to recruit interviewees automatically
- Use tools to align interviewers with interviewees
- Create a running product discovery backlog
- Establish a regular schedule for reviewing customer feedback
- Reduce friction by standardizing feedback across channels
Create a process to recruit interviewees automatically
The process of recruiting and scheduling customers to interview can be a hassle, as we saw in the quick example laid out above.
But what if you already have a qualified list of customers that have indicated their interest in participating in future interviewees? The dream!
In the past, maintaining a list of interviewees that could be sourced from has worked wonders for me.
In my previous organizations, to automatically have that list build itself over time, I’ve used a plug-in called Hotjar.
It is easy to set up, for example on a customer dashboard, and configure it up to automatically ask people logging in if they would be interested in participating in user research in the future. There are loads of easy to implement web plug-in’s that can do this as well.
If they selected ‘yes’, they would be prompted with a follow-up question regarding what types of research they would be interested in – surveys or customer interviews, for example – and an additional question or two to get more specific information.
If your volume of customer of logins is high, you can limit the questionnaire exposure, for example, to 20% of users logging in each day. Also, make sure you don’t ask people again if they’d already said ‘no’ – which is easy to manage via Hotjar settings. The result of this is a self-growing list of opted-in customers, along with additional metadata on the Hotjar platform.
Use a tool such as Zapier to automatically update a Google Sheet by pulling from the customer list in Hotjar. This Google Sheet represents the master list of customers, sourced across different channels, that had indicated they would be interested in participating. In addition to this, there are likely other channels to automate the recruitment of participants for product discovery from, such as:
- Native mobile apps
- Customer support channels, including email and chat
- Offline at customer events
If your team is using a platform such as Intercom, there are many built in-built tools you can use to embed into an email and chat correspondence to source and categorize participants from. Also, larger platforms such as intercom have Zapier integrations, so you can automatically update that Google Sheet with newly signed-up participants. Tools like Intercom are great because they automatically capture useful metadata for each user as well when set up correctly.
In the past, I’ve used Google Sheets – it doesn’t have to be a Google sheet – that was constantly being updated with new participants and useful metadata. When I talk about metadata, I refer to data points like:
- Source channel
- Type of research they want to participate in
- Demographics about them, such as user profile – you can pre-build these in many CRM platforms – products they are using, length of time as a customer and so on
- Contact details
- Additionally prompted question responses
This is important because it gives you another way of slicing and dicing the data. When you compare this against your total customer base by cohort, you get a more realistic picture if you can create relevant sample sizes for a particular cohort to interview.
Having these cohorts in place helps with the research. A lot of time can be spent creating these cohorts – having to sit and manually query data from different tables and stitch them together – if it doesn’t already exist.
This list might need to be partially manually maintained, specifically to note which customers have been interviewed in the past, filling out some metadata manually if it isn’t auto-populated, and so on. However, the time spent doing this regularly is trivial compared to setting up a new list from scratch each time.
Use tools to help match interviewers with interviewees
Ah, the joys of dealing with people and their clashing schedules. Make your life easier by using tools to help you do this. In terms of scheduling, there are 2 sides to schedule:
- The interviewee’s time
- The interviewers’ time – note: interviews should be led by at least 2 colleagues
It’s perfectly fine to manage this manually with a Google Sheet. In the past, I’ve used tools such as Calendly to make my life easier, but this is by no means essential.
As long as you maintain a sheet to coordinate schedules, statuses, participants, and updates, you’ll be fine. Unless you’re like me – terrible at this type of organization – hence using tools to do it for me.
Create a running product discovery backlog
Similar to how you would manage a delivery backlog, treat discovery in the same way. Collate ideas and feedback from customers and colleagues into one place. You can use tools such as JIRA, Trello, or even Excel for this.
The process of scoring and prioritizing items for discovery should be in line with your prioritization processes – more on this in an upcoming article.
Establish a regular schedule for reviewing customer feedback
To make use of feedback, it needs to be reviewed frequently. Whether it’s insights learned from an 8-week qualitative discovery initiative, or monthly feedback reviews, making a routine of these types of reviews with cross-team colleagues is important because it aligns everybody with up-to-date information.
One of my favorite review sessions is looking at the monthly feedback that gets raised from customer support teams. If you standardize your feedback appropriately, you can easily visualize changing trends.
For example, the frequency of different types of complaints raised to your customer support team or the number of different types of features or service requests can be significant. You’re getting into the quantitative world now, which can be used in parallel with qualitative approaches to help your organization respond effectively.
In terms of web analytics, I encourage you to use business intelligence tools to create dashboards and retrieve live data. For example, you could track onboarding funnel conversion rates or back-end events from your systems. This information is just as valuable as qualitative customer feedback because it clearly identifies where problems are occurring and how often.
Reduce friction by standardizing feedback across channels
This is such a big topic that it gets its own article later on. For now, there are two important aspects to focus on when standardizing feedback:
- The collection and aggregation process of feedback across channels
- Sanitizing, categorizing, and tagging feedback so it can be more easily analyzed when aggregated
The way you collect feedback is important. Standardizing the way you conduct interviews and take notes, how you categorize feedback, how you communicate findings and the tools you use across channels to gather feedback from customers are all important.
Standardizing these processes will make analyzing the feedback more efficient and insightful. Imagine – maybe you don’t have to imagine! – you have feedback coming from different channels, in multiple languages, in both structured and unstructured formats that get stored in multiple locations and different formats. Does that seem like something you can – or are motivated – to manage and dig into?
The goal is to have an aggregated view – in my case, it’s often simply a Google Sheet – of different user feedback per row with the relevant metadata to make it easy to analyze.
When you get that feedback aggregated, chances are that you need to do some tidying. This might require creating new fields manually to categorize different types of feedback into concise tags. For example, “Feature request for X” or “Complaint about Y service”, or perhaps creating different types of category themes.
Regardless, the time spent manually tagging and tidying up feedback in one place on a routine basis – once a month is feasible if you only deal with a few hundred points of feedback a month – is worth it versus having thousands of rows of unstructured and uncategorized feedback a few months later on that you’re trying to update in one batch. Grab a cup of coffee on the first Monday of the month and smash it out!
Unless feedback is standardized in a way that makes it useful, it can’t be used to its full potential. This can have consequences for your business in the form of missed opportunities and missed chances to address upcoming problems before they get really bad.
The good news though is that with a little bit of automation and routine maintenance, this becomes feasible and frees up more time for analysis. In an upcoming article, I will go into more detail about the individual steps in this process.
There is no golden methodology, but methodology is golden
There are hundreds of articles and books that exist today that go into detail about different research methodologies and how to apply them. The Nielsen Norman Group is a great place to start learning about different research methodologies. There are thousands of qualified professionals specializing in research. Google is your friend for finding out more.
I have a methodology that works for me – it’s crystallized over time into an 8-week discovery initiative that’s great when you have a specific target audience and objective in mind. More on this soon.
Having said that, there are many different approaches out there. So work with what skills your team has and go from there. There is no one golden rule to follow when it comes to methodologies of product discovery, other than to use a variety of them for different purposes and scenarios.
In God we trust – all others must bring data
To analyze all the different types of data across your business, you need to have some fundamentals in place. This doesn’t mean you need a fancy data team – though it’s great if you do! – but having a few people that can get their hands dirty writing queries, setting up business intelligence tools, and stitching these things together is key. So when it comes to your data, it’s good to ensure that:
- Analytics, especially web, is easily accessible via an intelligence tool that a non-engineer can use
- Quantitative and qualitative information is easily accessible
- It’s standardized
Going through the process of setting up custom events to track user journeys and plug-ins to gather feedback won’t give you insight unless you can easily analyze that data. This is where you should work closely with your data and engineering teams – if you have them – to align on what is important to track and how to pipe it together.
There is also no point in having great analytics if it’s not easily accessible by the business. Whether it’s a dashboard monitoring live onboarding funnel conversion data or findings from an initiative about a new target customer segment, if your information isn’t easily accessible then it won’t be used to its full potential. This can be solved by having a basic product discovery knowledge base in place, especially for qualitative insights.
Next up, how to create a product discovery knowledge base for your company.
Questions? Feedback? Get in touch.