You did your interviews and have your problem set out for you. But what do you need to focus on for building ventures? A problem has a million different subtle details. How to prioritise these?
Last week we talked about solving the right problem by conducting your interviews. This week, we'll use data to find the most important aspect of your problem. For this, you need more than five responses. You need quantitative verification. We do this verification in two steps:
- Gather your data
- Analyse your data
1. Gathering data
First off, gathering your data. Like last week, we split this up in B2C and B2B. Each type has a different way to get your data.
For B2C, the best way to get quantitative information is by doing a survey. Surveys are (quite) objective and allow you to compare different solutions and preferences. There are a couple of ways to set them up, and then you also need to get them filled in.
One way to set them up, is by listing a number of desired outcomes. People should rank these outcomes on satisfaction and importance. For example, for Uber, one of the desired outcomes might be 'Minimising the distance to get to the vehicle'. In your survey, you ask 'How important is this to you (scale 1-5)'and 'How satisfied are you with current solutions (scale 1-5)'. Uber solves this by bringing the car to you, but busses don't for example. More on the analysis later.
The benefit of asking for importance and satisfaction is the consistency and the objectivity. By asking all the questions in the same way, it's easy to compare and analyse. On the other hand you need to ask a large number of desired outcomes (50-100 for completeness). On top of that the way of asking and the number of questions make the survey boring. This makes it hard to get a large number of responses or makes it expensive if you outsource it.
Another way of doing surveys, is by performing a lot of small ones. These smalls surveys are easier to get. Also, given enough of them, we mostly have the same information as the previous method. We recommend 5-10 questions per survey. There are some new and trendy ways to get these filled in, like Google Opinion Rewards and Buffl (in Belgium). A negative is that you need more of those surveys. And because not everyone answers every question, you can't link every answer.
In a couple of surveys around spending on holidays, we found a huge difference between 25-27 year olds and 18-21 year old. Against our expectations, the younger folks used pen and paper to repay their friends after a week of holidays. The 25-27 year-olds used apps much more, specifically Splitwise.
Lastly, you need enough answers for statistical relevance. For the larger survey, we recommend at least 150, best to have more. For the smaller ones, you can have smaller number for each smaller survey, but you need to have more surveys. Do this again, in a growth-hackery way. Get connected to communities, go to events, promote your survey. An idea to do as well is tit-for-tats. We provide this benefit for your community, you fill out the survey.
For B2B, getting a large number of respondents is even harder. Your number of users is smaller and they're busy. Still, getting quantitative insights helps a lot for setting priorities. There's a couple of ways to still get this done.
First of all, you can do a survey, aimed at professionals. Here, it's best to stick with the JTBD survey style. You just ask for importance and satisfaction. This is because it's harder to do follow up surveys so you can't use the smaller one. Keep it slick, professional, clear and to the point. If you want the same number of results, B2B will be more expensive than B2C.
A second way to go quant, is to have more interviews. Keep these a bit short and focus on getting a real sense on the problems. Ask the participants to rate problems on a scale so you don't need to guestimate the need. This also allows you to compare your interviews better. And like the last suggestion, document a lot. Don't think you will remember what the first interviewee said 15 interviews down the line.
The last way is to use surveys that were already done. There are a lot of firms getting market insights and priorities. They won't ask the exact question you need to be answered. But they can give you interesting information. Some examples are Gartner, CB Insights, the million consultancies out there, universities etc. You can even check out your competition. They often have interesting information on their priorities. It's best to use these extra surveys in combination with the other methods.
2. Analysing data
Gathering the data itself is not the interesting part. The goal is to analyse it. Depending on the way you gathered your data, there are different ways to analyse it. In this section, we'll have a look at different methods and some points of interest.
Market Basket Analysis
If you did larger surveys, you can use market basket analysis. This is a fancy term to say, 'What are problems that occur together often?'. We say something is a problem when it's important but not satisfied. This gives you a very nice method to prioritise your issues and which needs to tackle together. Our tool Unveil help you analyse the long survey style of data.
Smaller surveys need another way of processing. Not everyone answered all questions, so it's harder to connect the dots. We set out key takeaways. And we quantify the questions for each survey. This way, we can compare needs between takeaways to solve the most important problems. You can also follow up the key take-aways with additional research
Lastly, select your problem well. Don't always just take the biggest issue. Look at the combination of problems. Look for niches that have a real problem that you can solve. Solving a big problem for a small community is better than marginally improving a small problem for a larger community. Look at the team and know your strengths. Don't make a car if you only have software engineers.
After the interviews and the analysis, you know exactly which problem to tackle, and which aspect to focus on. You have a firm grasp on the problem from the interviews. you know how to prepare the interviews, how to find people and how to process them. You also know how to gather data and how to process them, both for B2C and B2B.
However, don't overestimate this part. You need to also solve the problem, not just understand it. In your design sprint, you will again ask for feedback. There's always more room to pivot your solution to another problem. As we would say 'Nothing ventured, is nothing gained!'. And the (ad)venture doesn't end here. The best part is still in front of again.
Stay tuned for our next blog post in the series, about the design sprint. In one magical week, you'll set your goal, create a design, craft a prototype and get customer feedback. You'll be building ventures in no time!