PivotTables part 2: Creating summary reports made easy

One of the goals we had for our PivotTable work in Excel 12 was to make creating PivotTables a much more approachable task.  In this post, I would like to walk through an example of creating a PivotTable in order to highlight the changes we have made in Excel 12.  In general, we tried to make the process simpler and more intuitive.

As I said in the previous post, PivotTables are great for summarizing large amounts of data.  For example, a user might have a table full of sales data (contained in a query table, that they copied from elsewhere, that they have been typing into the grid over time, etc.) that contains products, sales figures, product categories, etc., and might want to see a summary of sales grouped by product and product category.  This is exactly the sort of thing that is easy to create with a PivotTable.  To start, the user needs to tell Excel they want to create a PivotTable.  There are two places they can do this.  First, on the Insert tab, where the first button in the ribbon is an inset PivotTable button.

(Click to enlarge)

Second, on the Table tab (the tab that shows up when I am working with a table of data), we have added a command to “Summarize With Pivot”.  This is essentially the same command, except when you use this command, we know that you want the table you are working with to be the data source for the PivotTable.

(Click to enlarge)

Once the user selects one of these commands, they are presented with a new dialog for creating PivotTables.  This dialog replaces the existing multistep wizard with a simple dialog that only presents the user with the most necessary choices … our usability research showed that a lot of users never made it past the wizard due to the complexity of choices required.  (Note, there are probably some out there wondering about whether they can still pivot against multiple consolidation ranges, etc. … the answer is yes.  The existing wizard is still available for advanced users that want to take advantage of its functionality; it just isn’t the mainline UI):

In our example, because we had a table selected when we created the PivotTable, all our user has to do is to click OK and the PivotTable is created – that’s a total of two clicks.  If the user wanted to work with data external to Excel, they would simply select “Use an external data source” and select the external data connection they would like to use from the drop down.  (Note that in the Table/Range refedit control, you can now type “structured references” to tables or parts of tables in addition to cell references.  This is true for most of the refedit controls in Excel, and is another small but powerul addition to the product.  See this post for more on structured referencing.)

After clicking OK in the dialog, the user sees a new PivotTable in the Excel grid, and a field list (significantly changed from previous versions) with which to populate the PivotTable:

(Click to enlarge)

Now that the PivotTable has been created, the next step is to add the data the user wants summarized.  This is another area where our customer research and usability studies demonstrated that many users had trouble, specifically in three areas.

  • Figuring out that they needed to get their “fields” onto the various areas of the PivotTable
  • Deciding which area, or “drop zone” of the PivotTable they needed to add their data to build their report.
  • Figuring out how to get the field to the drop zone (drag-drop not being an action that is all that common in Office applications)

In order to address these problems, we made a couple of changes.  First, we added checkboxes to the field list, which, though a simple change, clearly advertises to the user what actions they need to take.  Second, we changed the rules, so that users don’t need to move fields to drop zones to get going – when they check a checkbox for a field, that field is automatically added to the PivotTable, using a simple but generally effective algorithm.  If the data type of the field is numeric, the field is added to the Values area, with a default aggregation of sum. If the data type is non-numeric, the field is added to the Row Labels area.  As you click more non-numeric fields, Excel places them on the inside of fields already on the PivotTable, building a hierarchy.  Un-checking a checkbox will remove the field from the PivotTable.

So with the new field list, a user can get a quick summary by simply checking a couple of checkboxes, and that’s it.  Let’s walk through what this looks like.  As I said, our user wants to build a sales summary for products by their product category. The first field for the user to add is Product Category, so they click in the checkbox for that field …

(Click to enlarge)

… and the items of the Product Category field are immediately added to the PivotTable.

(Click to enlarge)

Next, to show the individual products of each product category, the user adds the Product Name field:

(Click to enlarge)

Now we have the products nicely listed under their product category in the PivotTable, complete with some UI that hints that you can expand and collapse levels … more on that in a later post.

(Click to enlarge)

And finally, our user adds the Sales Amount field to finish their summary report.

(Click to enlarge)

They probably want to format it as a currency too, which is one click on the ribbon.  And, with that, our user now has their summary report – Excel has grouped all the original data by Category and Product, calculated the Sales Amount, and slapped on a Grand Total … all in a total of 5 clicks.

(Click to enlarge)

I can hear some of your saying “but I liked dragging things around”.  I can also hear some other folks asking “how do I get things to the Filter area or the Column area to build a crosstab?”  (And probably a lot of other things, but I will stop guessing, address those two, and hear the rest in your comments.)  Well, you probably noticed in the screenshots above that as I added fields to the PivotTable, they appeared in the lower section of the field list. This section holds the drop zones, which were designed around two key points – making it easy to determine the current placement of fields in the PivotTable, as well as making it easy to rearrange the fields in a PivotTable. There are four areas in a PivotTable, each of which is represented by a drop zone in the field list.

  1. Report Filter. This area holds the fields that the whole PivotTable is filtered by.
  2. Row Labels. This area holds fields that act as labels for the values, and the labels appear to the left of the values.
  3. Column Labels. This area is just like row area, but the labels appear above the values, breaking them out by column instead of row.
  4. Values. This area holds the fields that are summarized (for example sales amount). Fields in this area are typically numeric, but can also be non-numeric (in which case they are counted).

If you prefer to drag fields from the field list to the drop zones, you can … this is also the answer to how you add fields to the Filter or Column drop zones.  To change the report layout, you can drag and drop fields between the different drop zones or you can click a field in a drop zone and select which area to move it to in the menu that pops up.

And finally, for those of you that really want the drop zones in the grid, we have put in a toggle to bring them back, but there are some additional considerations to that one, so I will cover it further in a later post.

So, to sum up, we have worked to make PivotTables easier and faster to create, and our usability testing with both beginning users as well as PivotTable experts (see post on usability studies for more on that subject) show that both user groups benefit from the new design.

Next week, much more on PivotTables.