The "I" in BI is "intelligence". Not "insight".
The BI Pipeline
Data is the new petroleum. Analytics is the refinery. Insight is the gasoline. Intelligence is the power that thrusts us forward. In this context, data is the raw material. Analytics is merely the processing; itself has no more intrinsic value than the data products it produces.
Our APG is an example of insights-centric data product. In BI, there has to be more, however. In particular, even though "actionable insights" is a popular term and aha moments are really nice, decision makers need much more in their decision making space.
Intelligence is different from insights in some fundamental ways. Insights can come in bits and pieces, can be fragmented, can take a narrow scope, and can even be contradictory, especially when insights could be taken out of context and generalized beyond their limits. In contrast, intelligence requires the process to gauge relevance, priority, and validity, as well as a methodology to integrate insights into a coherent whole.
A "finished" intelligence has to be summarized, clear, and, most importantly, targeted for the high-valued questions that are asked. Based on our experiences working with many corporate users, the most valued question-set looks like the following.
- How are we at [X] overall?
- How do things break down across ____ ?
- What's "the normal" vs. "the abnormal"?
- What are our priorities?
Answers to the above questions require the creation of a highly accessible decision making space, and fundamentally, an ability for decision makers to SEE all of the complex relationships in the decision making space. In other words, a useful intelligence product should show "the big picture", all of "the key specifics", as well as relationships between the big picture and the specifics.
The structure of the above question-set is very general too. For example, in product management, [X] can be "meeting our customer's most pressing needs"; in employee engagement, [X] can be "providing the most productive work environment"; in B2B relationship management, [X] can be "communicating with or servicing our partners"; in sales, [X] can be "how do we compare with our competitors" or "what drives purchasing decisions"; and in the public sector, [X] can be "what are the strengths / weaknesses of our community" or "what investments to make in our school system", ...
Creating Complete Profiles
Survature technology is about answering questions with data. If you don't want to capture the decision making space for the above question-set, without having to create power-points after power-points, please use Complete Profile.
The process is intuitive, real-time, completely DIY. You need to have created segments for the survey already (segment builder help-d0c). We recommend that you create a Date Range based segment first, choose the range that's from start day and end day of the survey. Name that segment as "All", for example. For the breakdowns, you can be flexible, any existing segments can work, per organizational division, geographic areas, product lines, demographics, purchase history, etc. can all work.
In your crosstab, pick "All" as the row selection, and pick the breakdown segments for the column selection (for instance, by geographic areas, such as East Coast vs. Midwest, vs. West Coast). The crosstab looks like the following.
The "EXPORT PROFILE" button will bring up a configuration menu of the profile (below). You can pick which AnswerCloud questions to include, also toggle whether to highlight niche priorities (in contrast to general priorities). When breaking down segments into micro-segments, it is often useful to know the population size. You can do so using the "Display Response Counts" toggle.
Given the widespread popularity of Net Promoter style questions, there are many scenarios where an 11-point rating question is asked but is really not about net promoter score. In that case, feel free to toggle on "Display NPS Average" to show the rating as averages.
After you have the configuration set, click on "OK" will generate the complete file for you. For large surveys or ones with complex segmentation criteria, the auto-generation could take up to 3 to 5 seconds.
Please note, the row selection doesn't have to be "All". If you need to ask the question-set on the basis of just "Region: East", then pick "Region: East" as the row, and pick employee promoters and detractors as the columns. In this way, our users can have the power of flexible deep dives, by creating complete profiles for the company as a whole, for an individual division, for an individual product line, etc.
Using Complete Profile
For most users we have seen, each complete profile they use for their strategy development can fit onto a single page.
When reading a complete profile, please start with the header-row, where each segment is identified. If a NPS score question was chosen to display, the score is shown right next to each segment. Depending on the toggle configuration, that score can be shown as an average rating or in the classic NPS format.
Below the header-row, each aspect of your decision making space is shown as individual AnswerCloud questions. The question's original text and number of responses are shown first, under which only items with the four (4) highest priorities are shown.
Four is the top 1/3 of all of the answer items, assuming your AnswerCloud question has a dozen answer items. If less than four (4) answer items land in the top-half of the two dimensional priority chart, then only those items in the top-half priority space are shown.
Beside each answer item, a 5-point average rating is displayed. That data is by averaging the explicit answers of each question. The interpretation of that rating is specific to the question asked. For example, "satisfaction with feature" is obviously different from "frequency of pain point".
Normal vs. Abnormal
- If you picked "Highlight niche priorities", all answer items highlighted in light pastel yellow are "niche priorities". In other words, when you analyze "how do things breakdown over the columns", these "niche priorities" appear in one and only one column. All of the unhighlighted items are general across multiple columns. In this case, you can tell what's "the normal" vs. "the abnormal" on prioritization.
- Another aspect to compare normal vs. abnormal is to compare the 5-point average ratings. In general, anything under 2.0 or above 4.0 should be considered as "abnormal". Of course, satisfaction scores of 4.8 would be great, 1.9 would be bad. However, the exact thresholds to apply would have to be different from one corporate setting to another.
Lastly, while a real-world decision making space will always have a high dimensionality, different aspects of the space are related, thankfully. Reading each column in a complete profile vertically is crucial to understand all of the nuanced aspects support and relate to each other. In the example used in this help doc, Q4 "individual needs", Q5 "team needs", Q6 "supervisor support" are intrinsically interrelated.
Similarly, if you are needing to understand your client's "total ownership experience" of your product, service, design, etc., you should expect to find relationships between purchase driver, product expectation, new product experience, long-term experience, market influence, etc. to all the related too. A complete profile can help you distill and integrate all of these insights into a connected whole (and onto a single page!).
To this end, please note that getting the most powerful complete profiles would have to start with an optimized design. Lazy design doesn't work. If you have specific new project needs, please talk to our user support!