Tutorial 5: Analyzing Your Data
Once you've closed all your missions, it's time to start your analysis. Follow these best practices to ensure you capture as many insights from your data as possible. No matter what kind of project you're running on dscout, you'll want to follow step 1 below. And if you're considering building a large tag list, reading steps 2-5 will make your life much easier when drawing conclusions.
The dscout approach to analysis
- Immerse yourself in entries individually, in aggregate, and offline
- Develop a tag list that identifies descriptive and thematic trends
- Build and review your tag list in dscout
- Evaluate and adjust your tag list, then tag every entry
- Examine relationships, quantify themes, and uncover insights
1. Immerse yourself in individual entries and high level patterns
Before you start organizing, immerse yourself in the data to get familiar with the stories you’ve collected and discuss what you find with colleagues. What's interesting or surprising about what scouts are showing and telling you? First review some entries individually, then absorb some of your data in aggregate to help surface initial findings and inform a more strategic approach when you go a step deeper with your analysis.
View individual entries
Click on an entry to bring up an expanded view. Here you can read responses, view media and see scout information. You can also add a bookmark to flag an individual entry with a single click, or write a note to share your opinion about a particular scout's entry with your colleagues.
View entry data in aggregate
The “Entries” dashboard automatically aggregates and organizes all of your data as it is submitted. Just click into the Part name or other headings in the left sidebar to filter by closed-ended question responses, demographics, scout groups & rankings, tags, or bookmarks. You can then investigate the entries that meet your filter parameters.
You can also click into the summary view (by either clicking the Entries dropdown in the top utility bar, or the Summary icon in the left sidebar) to see how your selected segment's responses compare to all your participants' responses.
Discuss and sort key stories
Sharing your initial impressions with colleagues at this early stage will help you identify and craft the story your data is telling you. Try printing out entries from your favorite scouts and sorting them into categories. Or ask each person on your team to read entries from 4-5 scouts and report back any interesting findings. Activities like this will help you gain a robust understanding of your entry data and make you aware of individual examples that complement your quantitative findings. Just click the arrow icon in the bottom right corner to open your export options.
2. Developing descriptive and thematic tag lists
Tags are short, descriptive labels that can be attached to individual entries. If your data is highly qualitative, it may not be necessary to quantify your data using tags. On the other hand, if you've packed your entry script with closed ended questions, you might be able to filter and drill down into your data without needing to use tags. But if your research project is complex, tags can be an essential tool for categorizing, quantifying and sorting entries with attributes you've defined.
The following example is taken from a dscout mission called "Saving Moments," which asked scouts to record instances where they felt like they were saving money. Scouts captured a huge variety of moments, from using 2 for 1 coupons, to avoiding impulse purchases, to using a mobile app to track their spending. Here's how we started making sense of all that data:
First we came up with some descriptive tags to mark objective, relevant things that we could observe in the entries. Descriptive tags are often situational (locations, actions, triggers) and can be grouped according to category. Keep in mind, there will be tons of things you can tag descriptively, so be sure to focus on areas that are likely to be relevant to your research question. For example, for the "Saving Moments" mission, we created just two kinds of descriptive tags: the tool scouts used to save money (like an app or website), and an action that described what that tool helped do (like avoid spending or investing).
Next, we came up with some thematic tags to quantify deeper themes that started bubbling up after immersing ourselves in the data. And after identifying the actions taken and the tools used to save money, it made sense to start tagging implied reasons for why scouts engaged in that behavior (like the desire to grow wealth or reduce the cost of essential purchases).
That effort, in turn, highlighted the fact that scouts were sharing saving moments that could be grouped into three main categories–moments where they lost money, moments where they gained money, and moments where no money actually exchanged hands. This led us to modify our thematic tag list and group them into the goal and cash flow tag groups below.
Developing a smart thematic tag list takes time, so don't be discouraged if you don't see a whole bunch of connections right away. The longer you sit with your data, the easier it will be to use thematic tags to quantify things like pain points, stages of a journey, different modes of behavior, or competing strategies.
3. Build and review your tag list in dscout
Once you've built your tag list using a chart like the one above (or sticky notes!) it's time to build your tag list inside dscout. Just make sure your tag list is fully mapped out first. Otherwise, you'll end up with a mashup of inconsistent tags. Here's how to get started:
- Page through your entries to view each in detail (you can use your keyboard's arrows if you'd like), adding tags and creating tag groups as you go using the "tags" tab on the right side of each entry.
- After creating a new tag, you can choose which tag group it goes in, create a new tag group, or enabling auto-tagging by clicking the "A" icon to the tag's right.
- In the example below, the scout has submitted a entry about buying two pairs of glasses for the price of one. Based on his responses, we've tagged "action: coupon, deal" and "cashflow: negative". (You'll notice there's an "other" tag in each group, which is checked if none of the listed tags fit a entry. This really helps keep track of which entries have been tagged completely.)
4. Evaluate your tag list, then tag every entry
- Does it feel difficult to apply your tags? A good tag list should feel easy to apply and maintain. If lots of tags seem like they could go either way with an entry, or if too many entries seem like they could fit into any tag, it's time to review your tags and your approach.
- Cut your thematic tags by your descriptive tags and other demographic attributes to see if you find distinctions. If there are no noticeable differences between the themes, they may not be meaningful themes to focus your analysis on.
- Is your tag list mutually exclusive and comprehensively exhaustive (MECE)? This is ideal, but not always possible. Strong knowledge of your data will help you know when to strive for MECE tags and when it would complicate your analysis.
- Are there any tags that overwhelm the rest of a data, such as a tag that could be applied to 40% of your entries? If there are, try to break that tag down into smaller parts.
- Are there a bunch of entries tagged "other"? If so, revisit your "other" entries and look for commonalities. Can you create a new tag for some of these entries to call home?
When you're confident that your tag list is right for your data set, dive back in and apply tags to every entry.
5. Examine relationships and quantify themes
When all your entries are tagged, you can turn to dscout's sorting, filtering and export features to make sense of the data. Up to this point, you've spent your time identifying interesting themes and categorizing your data points with attributes that inform these themes. Now's the time to quantify how often these attributes occur and examine relationships between them.
Start by examining the frequency of an attribute within the entire data set.
- Out of all the saving moments, how often were scouts utilizing an app? (tag = tool: app)
- Out of all the saving moments, how often were scouts working to prevent purchase? (tag = goal: prevent purchase)
- Out of all the saving moments, how often is the scout actually spending more money than they're saving? (tag = cashflow: negative)
Then examine the intersections between attributes (the frequency with which combinations occur).
- When the goal is to prevent a purchase, which saving tool is most frequently mentioned?
- When the saving tool is an app, what goal is most frequently mentioned?
- With people 18-34, which saving tool is most frequently mentioned?
- Do men or women more frequently use coupons to save money?
To answer these questions, we'll segment the data by thematic tags, descriptive tags, demographics and responses to questions. You can apply multiple filters to see the frequencies of a specific subset of entries. In the example below, you can click the tag "tablet" to only display the entries where a coupon was used. Then, using the demographics tab, you can view only entries wherein a woman used her tablet to view media.
If you're looking for more rigorous quantitative analysis, you'll likely want to use crosstabs in addition to filters. A crosstab (short for cross-tabulation), is a table that displays frequency counts for your data set. Crosstabs allow you to evaluate the intersections of certain attributes, by displaying the frequency with which combinations occur. Click the export icon to create a crosstab in .csv, with the options to include tags, questions, scout groups, age, gender and more.
Once you've wrapped up analysis, it's time to tell a story. At dscout, we generally deliver our insights in a framework. Your insights are the things you learn, your framework is the way you relate those things together, and the story is the way you communicate the relationship between the two.