Creating Looks (and Working with Explores)

Looks are an integral part of Customer Insights: not only are Looks useful in their own right but Looks also serve as a key building block for Dashboards.

With that in mind, you might be thinking, “Wow, I better find the Create New Look button and start building some Looks.” But here’s a tip before you do that: there is no Create New Look button in Customer Insights (nor is there any sort of Create New Look menu option). In fact, there’s no way to “create” a Look: instead, you start with an Explore, and then save that Explore as a Look. That might sound a little complicated but, as you’re about to see, it’s actually pretty easy.

And no, you can’t “save” an Explore: Explores are (depending on how you want to look at it) read-only. (Why? Well, for one thing, that way you’ll always know exactly what getting each time you open an Explore.) Just remember this formula: you start with an Explore, and then save that Explore as a Look. That might not be everything you need to know, but it’s a good start.

Note, too that, after a Look has been created, you edit the Look itself: you don’t make changes to an Explore. However, the editing pane for a Look is pretty much indistinguishable from the editing pane used by an Explore:

The Look Editing Pane

Because of those similarities, this documentation will primarily talk about working with Explores; however, pretty much everything we talk about is also true when it comes to editing Looks. The main difference? After editing a Look, you have to click Save to save those changes or Cancel to discard your changes:

The Look Save Button

Note. You also won’t find any Save As options: when you edit a Look, the only thing you can do is update that particular Look.

Explores don’t have a Save button. To save any work you’ve done in an Explore, you have to click the Options icon and choose either Save as Look or Save to Dashboard:

Saving an Explore as a Look

This section of the documentation includes the following topics:

  • The Explore Options icon
  • Adding Dimensions, Measures, Pivots, and Filters
    • Adding Dimensions
    • Adding Measures
    • Adding Filters
    • Adding Pivots
    • Adding Row and Column Totals
    • Adding Table Calculations
    • Modifying Row and Column Limits
  • Selecting a Visualization

The Explore Options icon

Each Explore includes a Options icon located in the right-hand corner of the Explore window. Click that icon, and you’ll see a series of options similar these:

The Explore Options Menu

These options are described in detail in the next few sections of this documentation.

Save as a Look

Enables you to, well, save the Explore as a Look. Bear in mind that you can only save the Look to spaces where you have the Manage Access, Edit permission; typically that will be your personal space as well as any Custom Reports space you have been granted access to. If you try to save a Look to a space where you don’t have this permission, you’ll receive a warning to that effect, and the Save & View Look button will be unavailable:

Permission Denied to Save a Look

The title you give your look must be unique within the Space where it is being saved. However, that title does not have to be unique across Customer Insights: you could have a Look called Demographics in every Space in your organization. (That said, however, giving Looks unique names might be something of a best practice.) Title, by the way are case-insensitive: Demographics is the same title as demographics and the same title as DEMOGRAPHICS.

Save to Dashboard

Enables you to save to the Explore directly to a Dashboard. When you do this, the Explore is saved as a Query Tile and not as a Look. That’s important to note, because Query Tiles are bound to a specific Dashboard. By contrast, Looks can be added to multiple Dashboards, and can be viewed even if they are not part of a Dashboard.

As you might expect, you can only save Looks to Dashboards where you have the Manage Access, Edit permission. If you try to save the Look to a Dashboard where you don’t have this permission the Save to Dashboard button will be unavailable:

Permission Denied to Save a Look to a Dashboard

Note that tile titles do not have to be unique, even on the same Dashboard; it’s not necessarily recommended, but every single tile on your Dashboard can have the exact same title:

All Tiles Have the Same Title

Download

Enables you to download a copy of the data retrieved by the Explore. For more information, see Downloading Data.

Send

Enables you to email a copy of the data retrieved by the Explore. For more information, see Sending Data.

Save & Schedule

Enables you to schedule delivery of the data (via email). For more information, see Scheduling Data Delivery.

As you probably noticed, this menu option is labeled Save & Schedule. That’s because you need to save the Explore as a Look before you can schedule it. Why? Well, as we mentioned previously, Explores can’t be saved: you can make as many changes to the Explore as you want, but the moment you leave the page all those changes disappear. You can’t schedule delivery of the data because the Explore that retrieves that data will no longer exist. Therefore, save the Explore as a Look, then schedule data delivery from the Look.

If you click Save & Schedule, you’ll first be shown the Save Look dialog box:

The Save Look Dialog Box

After the Look has been saved, Customer Insights opens the new Look and displays the scheduling dialog box:

The Scheduling Dialog Box

Share

Provides two different URLs that can be shared with others, and enable you to view an Explore configured exactly as the Explore you are working on (this is true even if you exit the Explore without saving it as a Look or to a Dashboard). When you click Share, you’ll see the Share URLs dialog box:

The Share URLs Dialog Box

The Short URL is a simple version of the URL. You can give someone the short URL and, assuming they have a Customer Insights account, they can use that URL to access a similarly-configured Explore.

Users can also access this Explore by using the Expanded URL; the only difference between the two URLs is the fact that the expanded URL includes the Explore details (including fields names and filters) as part of the path.

Note that these URLs never expire, meaning you can use them to re-create an Explore any time you want.

Remove Fields & Filters

Resets the Explore by removing all dimensions, measures, and filters, and by resetting the visualization to the default chart type (column chart). For example, suppose your Explore looks like this:

Removing Fields and Filters

After clicking Remove Fields & Filters, it will look like this:

Removing Fields and Filters

Clear Cache & Refresh

The first time you open a Look, the underlying query executes and the Look retrieves data directly from the database. A copy of that retrieved is then stored in the Customer Insights cache. If you run the same query a few minutes later, data will be returned from the cache rather than the database. You can verify that by looking at the Explore header; if it says from cache that means that data was retrieved from the cache rather than the database:

The From Cache Notification

Using the cache speeds up queries and improves overall performance. And, in many cases, you don’t need up-to-the-minute data anyway: for example, if you’re looking at historical data (say, logins for the previous month), the data is going to be the same regardless of where it comes from.

Note. We need to mention that data remains in the cache for one hour. If you run a query and then, two hours later, run that same query again, data will be retrieved from the database, and the cache will be updated. Likewise, data will be retrieved from the database if the query requires data not found in the cache. For example, suppose you run a query to return data for the year 2017; that 2017 data is cached. You now modify the Look so that it returns data for 2017 and 2018. Because the 2018 data isn’t available in the cache, the Look will directly query the database.

Of course, there might be times when you need up-to-date data. If that’s the case, click Clear Cache & Refresh. In turn, Customer Insights will empty the cache, and then query the database directly (with a copy of the new data being added to the cache).

Adding Dimensions, Measures, Pivots, and Filters

When working with Explores, there are several different terms you need to be familiar with; these terms include Dimensions, Measures, Pivots, and Filters.

Adding Dimensions

Dimensions are used to specify the data you want returned; more specifically, and using database terminology, dimensions return specific fields from specific tables. You use dimensions to determine the information you want your query to bring back. For example, if you’re using the Profile Dim Explore, you can return information about user ages, email domains, and genders. Those three items - Age, Email Domain, and Gender - are the dimensions for the Profile Dim Explore.

Adding a dimension to an Explore is remarkably easy. (In fact, sometimes it can be a little too easy: it’s possible accidentally add a dimension that you didn’t want to add.) To add a dimension, click the field name in the Explore pane:

Dimensions in the Explore Pane

Dimensions can be added either from the All Fields tab or from the Dimensions tab. The All Fields tab shows both dimensions and measures and, by default you need to expand each Dim and Fact in order to access those items. By comparison, the Dimensions tab doesn’t show any of the measures, but it does “pre-expand” each Dim and Fact:

The Dimensions Tab

You can also locate dimensions by using the Search field. For example, suppose you know that there’s a dimension named Au Format Date, but you aren’t sure where to find that dimension. In that case, type au in the Search field and then press ENTER. In turn, Customer Insights will show you all the fields that have the sting value au anywhere in their name:

Searching for Dimensions

Note. After searching for a field name, click the X to reset the fields list.

Once you’ve located the desired dimension, you can add it to your dataset simply by clicking the dimension name. When you do that, the background color of the dimension will turn to white, and the dimension will appear in the Data section:

Adding a Dimension

And yes, that is easy. The only tricky part is the fact that, when you run your mouse over a dimension, little buttons labeled Pivot and Filter appear. Clicking either of those buttons will not add the dimension to the Data section. Instead, and depending on the button you clicked, your Explore will use the dimension as either a pivot or a filter. If either of those buttons changes color, you know you goofed:

The Pivot Button

Fortunately, it’s easy to undo a mistake like this. Did you accidentally click the Pivot button, or did you inadvertently click the wrong dimension name? Don’t worry about: just click the button (or the dimension name) a second time and it will be removed from the query:

Removing a Pivot

In other words, click once to enable something, click a second time to disable it.

Note. You can also remove dimensions by hovering the mouse near the right end of the dimension, clicking the Setting icon, and then clicking Remove:

Removing a Dimension

To add additional dimensions, repeat this process as often as needed:

Adding More Dimensions

And what do you do after you finish adding dimensions? Well, one thing you can do is click Run and retrieve some data. For example, here’s the returned data from a simple little query that returns the names of all the countries in the database:

Sample Dataset

As noted, this query returns the different country names that are stored in our database. That can be useful information. However, what might really be useful would be to know how many users we have from each of these countries. That’s where measures enter the picture.

Adding Measures

Dimensions specify the data returned when you run your query. But what do you want to do with all that data? Once you have it In some cases, the answer might be: nothing. Simply gathering the data and being able to view and analyze it might be enough. In that case, you might not care about measures.

At other times, however, you might have specific questions you were hoping the data could answer. For example, you might want to know how many users registered during each of the last six months, or how many users you have from each country in South America, or maybe how many users are signing on by using a cell phone as opposed to signing on by using a PC.

Measures help answer these questions. The default Explores that ship with Customer Insights all include the Count measure, which simply tallies up the number of whatevers (a count of all the users registered during the last six months, a count of all the users from each country in South America, etc.). For much of the data that gets imported into Customer Insights the Count measure will suffice. For those instances where it doesn’t, you can create your own measures by adding table calculations to your query.

The best way to explain what Measures do is to show you what a dataset looks like with and without a Count. Here’s an Explore that includes a single dimension (Email Domain):

Email Domains

As you can see you, this is just a list of the email domain for each user profile in the database. If we displayed the entire dataset, we’d see a table with over 11 million rows. That’s … interesting … but it’s difficult to draw many conclusions from a table like that.

In this updated visualization, we’ve added the Count measure (we’ve also sorted the data in descending order):

Adding the Count Measure

Now we can draw some conclusions: gmail.com is the most-commonly used email domain in our database, followed by hotmail.com and yahoo.com.

To add this capability to a query, just click the appropriate Count measure. And yes, the word appropriate is important. If you are counting profile-related information then use the Count measure from the Profile Fact Explore. If you are counting event-related information, use the Count measure from the Event Fact Explore. Doing anything else will lead to unexpected (and usually incorrect) results.

Oh, and here’s a neat trick: you can create a query that uses just a Count measure. For example, this query contains only the Profile Fact Count:

A Measure-Only Query

That means that only a single data point comes back: the total number of user profiles in the database.

Adding Filters

To add a filter to an Explore, locate the dimension you want to filter on and then click Filter. When you do that, the filter is added to the Filters section:

The Filters Section

If you’re adding this filter so users can filter your dataset then you’re done: you don’t have to do anything else. (Except maybe add additional filters.) However, you might want to use this filter yourself, you might want to return a subset of the available records. In that case, you’ll need to do two more things:

  1. Specify the filter operator.
  2. Specify the target filter value.

The filter operator determines the relationship between the filter field and the target filter value. By default, a new filter uses the is equal to operator; that lets you create a filter like, say, “The event type is equal to signin.traditional.” Or, for the more visually-inclined:

A Completed Filter

But what if that isn’t correct? What if you wanted a filter that said, “The event type is not equal to signin.traditional?” That’s fine; just click the filter list and select a different operator:

Changing the Filter Operator

Note. For a complete list of filter operators and what they mean, see the Filter Operators Reference.

That means that you can easily create a filter that looks like this:

Changing the Filter Operator

As for the filter value, you can either type in the value (e.g., signin.traditional) or, in some cases, click in the filter value field and select from a dropdown list. For example, if you’re working with event types, you can click in the filter value field and select from a list of event types:

Selecting Filter Values

These (obviously) are the event types available to you. You can type in a different event type, but there’s not much point in that: it won’t return any data because you don’t have any records that match that “custom” target field value. (If you did have records that matched that value, the value would have shown up in the dropdown list. )

And yes, you can add additional filters to your Explore. For example, here we not only have a filter for event type, but we also have a filter for Month Name and Year Number. That lets us do things like query for all the social sign-in events that occurred in March, 2017:

Adding Multiple Filters to a Look

If you want to get rid of a filter, just click the X next to the filter you want to remove. For example, if you click the X next to Event Type, your Filters section will look like this:

Removing a Filter

In addition to filtering on more than one dimension at a time, you can also filter on more than one filter value. For example, suppose we want information from January 2018 and from March, 2018. To do that, we start by entering our first filter:

Adding a Filter Clause

When that’s done, click the + next to this filter. When you do that, a new filter clause is created:

A Filter with Two Clauses

To filter for January and March, just enter March as the second target filter value:

Filtering on Two Values

In turn, the only data we get back are event types from either January or March:

Filtered Dataset

Incidentally, we didn’t have to add an additional filter clause to get back this data. (Now we tell you.) Instead, after entering January as the target filter value, we could have clicked in the exact same field and then added March:

Filtering on Two Values in the Same Clause

That gives us two values in the same field:

Filtering on Two Values in the Same Clause

When you have two values in the same target filter field, Customer Insights automatically does an OR search: it searches for events from January or from March. In fact, take a peek at the SQL query generated for this Explore:

Sample SQL Query

The WHERE clause says it all:

(date_dim.month_name = 'January' OR  date_dim.month_name = 'March')

Before you ask, no, there’s no way to use a filter to limit the data returned by a Look but not make that filter available to users. For example, suppose you create a Look and you include a filter similar to the one shown above. When a user accesses this Look, he or she will also have access to the filter:

Filters are Available in Looks

You might have preferred that the user not be able to access the filter, but there’s no way to hide filters from users.

Note. Or at least not if the user directly access the Look. If the Look is on a Dashboard, however, any filters attached to the Look will not be available to users. To prevent users from filtering data, create a Query Tile and add it to a Dashboard instead of a creating a Look and adding that to a Dashboard.

On the flip side, you can add “empty” filters that you, as the Look creator, didn’t use, but that you provide as a service to others. For example, suppose you include Country as one of the fields in your dataset. You might not filter on Country at all. However, you can still include an empty Country filter in your Look:

An Empty Filter

Note. Using the is equal to operator and not specifying a filter value means, in this case, “Show me all the users, regardless of country.”

When a user views this Look, he or she can use the filter to explore the data in different ways. For example, suppose the user in question is only interested in Canadian residents. In that case, he or she only has to set the filter to Canada:

Filtering by Country

Adding Pivots

If you’re familiar with spreadsheet programs such as Microsoft Excel, then you’re probably familiar with pivot tables. Pivot tables provide an “extra” way of grouping data. For example, suppose you have a query that returns the number of login and registration events that took place each day of the week:

Events by Day of the Week

It’s difficult (if not downright impossible) to glance at this table and draw any conclusions. For example, which day of the week experienced the most registration and login events? Which day of the week experienced the fewest login and registration events? Without adding up all the event totals for each day, there’s no way of knowing.

In other words, we not only need to group events by event type, we also need to group events by day of the week. That’s what a pivot enables us to do:

A Pivoted Dataset

To add a pivot to a chart, find the dimension you want to pivot on and then click Pivot:

Adding a Pivot to a Query

The pivoted field appears in the upper row of your data table; in addition, the pivot columns are shown in light brown rather than blue:

A Pivot Column

To remove a pivot, click Pivot a second time:

Removing a Pivot

You should limit each query to a single pivot.

Note. You cannot create a pie chart or a single value visualization if you query includes a pivot. However, in place of the pie chart, you can create a donut multiples visualization.

Adding Row and Column Totals

Another useful way to supplement data (at least in data tables) is to add row and column totals. For example, consider the following table, which shows sign-ins for the past year:

Sample Dataset

Nice, but wouldn’t it be even nicer if we knew the total number of sign-ins for the year? You bet it would. And we can add that information just by going to the Data section and clicking Totals:

Adding Column Totals

Doing that adds column totals to our table, and to our visualization:

Dataset with Column Totals

Now consider this table:

Pivoted Table

Here we show quarterly values for several different event types. Wouldn’t it be nice to see yearly totals for each event type? Yes, it would. But it would also be nice to see the quarterly totals: the combined event totals for each quarter. And we can add that to our table by selecting Row Totals as well as Totals:

Adding Row Totals

Here’s the revised table:

Dataset with Row Totals

Adding Table Calculations

Customer Insights default Explores include the Count measure and (in two cases, anyway) the Count Distinct measure. Both of those measures are extremely useful in analyzing data and in constructing visualizations. However, there will likely be times when you want to do something besides counting up the number of Xs in a dataset. That’s where table calculations come in.

Table calculations (created by using a formula syntax similar to that used in Microsoft Excel) provide a way for you to add custom measures to a data table (and, by extension, to a visualization). For example, suppose you have a table similar to this one, a table that shows legacy sign-ins on a month-by-month basis:

Sample Dataset

Let’s further suppose that you had set a goal of 35,000 sign-ins per month (420,000 sign-ins for the year). From the table, it’s easy to see that you met the overall goal: you had 459,070 sign-ins during the year. But how often did you meet the monthly goal, and by how much did you meet (or fail to meet) that goal. That’s not quite as easy as to say, at least not without doing a little math.

Doing a little math is good; adding a table calculation is better. When you add a table calculation, you add a new column to the table; that column contains whatever data is derived from the table calculation. To see what that means in real life, complete the following procedure:

  1. In the Data section of your Look, click Calculations:

    Adding a Table Calculation
  2. In the Table Calculations dialog box, replace the title Calculation 1 with the title Percent Difference from Goal:

    The Table Calculations Dialog Box
  3. Click Default Formatting and then click Percent (0) 123,457%:

    Assigning a Formatting Style

    In case you’re wondering, we want to show the percent difference between the target value 30,000 and the actual number of sign-ins. We want the value specific as a percentage (and without decimal points), which is why we selected Percent (0) 123,457%.

  4. In the formula section (directly below the title) type $. You should see something similar to this:

    Adding a Table Calculations Expression

    This is the list of dimensions and measures that are used in your query. To determine the percent difference, we need to take the value of the Event Fact Count field and divide it by 35,000. In the dialog box, select Count ${event_fact.count}. The formula field will look like this:

    Adding a Table Calculations Expression
  5. Now complete the formula by adding** / 35000** and then click Save Table Calculations:

    Adding a Table Calculations Expression
  6. Your data table will immediately redraw itself to look like this:

    Dataset with a Table Calculations Column

Now it’s much easier to see whether we succeeded or failed to reach our goal, and by how much (good or bad) we deviated from that goal.

Tip. Want to make it even easier to see whether we succeeded or failed to reach our goal each month? In that case, add some conditional formatting to the table. Without detailing all the steps, in this visualization we’re asking Customer Insights to colorize any monthly totals that failed to the goal. In turn, we’ll get a visualization similar to the following:

Table Calculations with Conditional Formatting

Here’s another cool way to use table calculations. As you might recall, our original table shows sign-ins on a month-by-month basis:

Sample Dataset

But what if you wanted to know more about sign-in trends, what if you wanted to know how signups for a given month compared with signups for the previous month? That can be a very interesting question, and Customer Insights includes a nifty feature - offset - that helps you answer that question.

By default, table calculations compare values in the same row. In this case, however, we want to compare values in different rows; for example, we want to compare the total sign-ins in the June row with the total sign-ins in the May row. With offset, we can compare a row with the previous row (by setting the offset to 1) or with the next row (by setting the offset to -1). Here’s how we do that:

  1. In the Data section of your Look, click Table Calculations.
  2. In the Title field, type Month-to-Month Trends.
  3. Click the formatting list and then click Decimals (0) - 1,235.
  4. In the Table Calculations dialog box, type $.
  5. Select Count ${event_fact.count}.
  6. Type - o. and then click offset(Any,Number). Your formula will Look like this:

    Adding an Offset Table Calculation
  7. With the cursor inside the parentheses, type $ and then click Count ${event_fact.count}:

    Adding an Offset Table Calculation
  8. Type ,1 and then click Save Table Calculations:

    Adding an Offset Table Calculation

Here’s your new visualization:

Dataset with an Offset Table Calculation

Note. If you click Calculations, you’ll bring up the Table Calculations dialog box. From there, you can apply formatting, change the column title, etc.

Modifying Row and Column Limits

By default, queries return the first 500 records that meet the query requirements. That occasionally results in visualizations like this one:

Column Chart

Is that a useful visualization? Well, part of it is. But because the highest value is so high and most of the other values so low, a good part of the visualization is just empty space.

The easiest way to fix this problem is to limit the number of records (rows) returned by the query; as a result, that limits the number of columns displayed in the visualization. To limit the number of rows, go to the Data Section and type a new value in the Row Limit field. For example, this limits the number of returned rows to 7:

Assigning a Row Limit

That gives us a very different visualization:

A Row-Limited Column Chart

By the way, if you place a limit on rows or columns you might see a message similar to this when your run your query:

The Row Limit Reached Warning

This message tells you that there are more records that could have been returned but, due to the row limitation, were not returned.

If your query includes a pivot, you can place a similar limit on the number of columns displayed:

Assigning a Column Limit

By default, Customer Insights has a row limit of 5,000 and a suggested limit of 50 columns.

Selecting a Visualization

It should come as no surprise that visualizations are selected from the Visualization section. To choose a visualization, click the arrow next to Visualization to expand the section. When you do that, the background color of the Visualization header will turn black, and you’ll see little gray icons of the different visualization options appear on the header:

Selecting a Visualization

If you hover the mouse over the bar, the individual icons will “light up,” and a label tells you which visualization type the icon represents:

Selecting a Visualization

Note. The last icon in the list looks like this:

The Three Dots Icon

This isn’t actually a visualization type; instead, click the three dots to display additional options:

Additional Visualization Options

When you create a new Explore, the column chart icon is selected by default; however, you can change the visualization type at any time just by clicking a different icon. For example, suppose your visualization starts out looking like this:

Sample Column Chart

Click the pie chart icon, and your visualization will instantly change to this:

Switching to a Pie Chart

And before you ask, no, there’s no way to delete a visualization entirely from an Explore or a Look. However, if for some reason you really don’t want a visualization do this:

  1. Change the visualization type to Table.
  2. Go to the Data section, click the Options icon for each column in the table and then click Hide from Visualization:

    Hiding a Column from a Visualization

If you hide each column then your Visualization section will be blank:

A Blank Visualization

But, again, we aren’t really sure why you’d want to do this.

Note. True, you probably don’t want to hide all the columns in a visualization, but you might want to hide an occasional column. For example, suppose you have a table that reports the web browsers used to access your site:

Hiding a Table Column

In this case, you might only be interested in the percentages and not in the raw values; however, you need the raw values in order to calculate percentages. If so, then all you have to do is hide the Count column:

Hiding a Table Column

Although you can select any visualization type for any Explore, the visualization you can actually use vary depending on your dataset. For example, suppose you have an Explore that uses a pivot:

Sample Column Chart

If you want, you can easily change the chart type to Donut Multiples:

Switching to a Donut Multiples Chart

However, you can’t change to a pie chart:

Unable to Use a Pie Chart

That’s because the underlying dataset doesn’t work with a pie chart. Try switching to a map chart and you’ll get the same net result:

Unable to Use a Map Visualization

Regardless of the visualization you pick, that visualization comes with a number of formatting options. If you click the Edit icon on the right side of the chart, a menu similar to the following will appear:

Editing Visualization Options

Each chart type has own unique set of options; for more information, see our Visualizations Reference.

See Also

Scroll ↓