The broadest classification of products in our hierarchy. In many cases, these instances are similar to store departments from brick and mortar layouts (Dairy, Frozen, Floral, etc.)

Categories are a classification under Departments. The Category level starts to break up the Department, but does not name narrow product types. For this reason, Categories are more valuable for navigation purposes, rather than product selection. Examples of Categories within the Dairy department include Cheese and Yogurt.

Subcategories are a classification under Categories. The Subcategory level gets closer to actual product names (Greek Yogurt, Fruit on the Bottom Yogurt, etc.) Users will generally see more shared characteristics across Subcategories than across Categories or Departments.

Subsets are a classification under Subcategories, representing the narrowest classification of products in our hierarchy. These nodes have defined products that can be selected, such as Greek Yogurt - Cup. However, users may need to apply Facets to identify a more specific product (i.e. “is Organic” for Greek Yogurt - Cup).

We assign Facets to any of the four levels of the hierarchy or to individual products. This classification by attribute allows us to offer our customers the highest level of granularity in our data insights. We offer two types of facets: Binary and Ordinal. 

• Binary facets are characterized by facet value statements such as “Is Gluten Free,” “Color Red,” or “Plastic Construction.” Products or hierarchy levels are assigned values of “Never,” “Sometimes,” or “Always” for all relevant binary attributes.

• Ordinal facets are used for characteristics that are numeric, such as “Weight Capacity” and “Overall Length.” Values for “Weight Capacity” might be 100 lbs or 1 ton.

Intent tracks online consumer activity through human and AI classification, measuring interest in particular categories, subcategories, and individual products. Intent assigns a rank to every product we see on the Internet, based on the likelihood a viewer of a given page is to purchase a specific product. This metric captures a pre-purchase picture of sales trends, offering insight before sales data is available.

Total Intent is calculated as follows:
Total Intent = (Page Views) * (Entity Intent)

• On a given URL, a product may be mentioned several times. Each mention is referred to as an Instance.

• Each Instance is assigned an Intent score (from -5 to 5).

• A product’s overall Intent score for a given URL is calculated by averaging the Intent scores for all Instances on that page. We call this grouping a Product Entity, meaning several Instances referring to the same product.

• Total Intent is then calculated using the formula above.

An easy way to think about this is by taking a hypothetical example of a recipe on abc.com. There are 3 mentions of Organic Skinless Chicken Breast that each receive +4 Intent. We group the 3 Instances into 1 Entity. Say this recipe on abc.com has been viewed 100k times in the last month. This means that abc.com would contribute +400k Intent for Organic Skinless Chicken Breast last month.

Content that indicates a likelihood of consumer purchase. Positive Intent results from web pages that demonstrate a favorability toward the indicated product. [See “Intent” for a detailed description of how a product’s Intent score is calculated on a given URL.]

An Intent Event refers to a product being viewed on a URL—a Positive Event occurs when a product is represented favorably, indicating a likelihood of purchase.

Content that indicates an unlikelihood of consumer purchase. Negative Intent results from web pages that demonstrate a negativity toward the indicated product. [See “Intent” for a detailed description of how a product’s Intent score is calculated on a given URL.]

An Intent Event refers to when a product is viewed on a URL—a Negative Event occurs when a product is represented unfavorably, indicating an unlikelihood of purchase.

The sum of all Intent Events within product criteria selected by user. This feature helps clarify the volume of Events that contribute to an Average Intent score. 

Average Intent refers to the average Entity scores for a given product, across all web traffic we process. For example, if we see 1,000 URLs that mention Butter, this metric reflects the sum of all those Entity scores divided by 1,000.

Average Intent = (Sum of all Entity scores) / (Total Entities)

[See “Intent” for a detailed description of how a product’s Intent score is calculated on a given URL.]

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Explicit product rejection. 
For example, an article calling “[Specific Brand’s] Full-Fat Butter the worst thing you can put in your body, worse than any other butter.”

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Specific product negativity. 
For example, an article claiming high-fat butter leads to high cholesterol.

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

General product negativity. 
For example, an environment-friendly blog post claiming that all animal products are bad (meat, milk, butter, etc.)

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Product avoidance. 
For example, an article advising readers to “avoid butter if possible.”

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Casual product disapproval. 
For example, a recipe that calls for “some type of greasing mechanism such as olive oil; butter is OK if that’s all you have.”

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Passively suggested in recipe. 
For example, a recipe that calls for “some type of greasing mechanism, such as butter, olive oil, or vegetable oil.”

Indifferent listing. 
For example, a recipe for “grandma’s homemade oatmeal raisin cookies” that requires a baking sheet shows awareness of, and the potential need for, a baking sheet.

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Casual product reference.
For example, a professional (e.g. Dr. Oz), blogger (e.g. Recipe Girl), or influencer (e.g. @spoonforkbacon), writing a post in which they casually make a reference like: “I like to use butter when I make this dish” or “Everything’s better with butter.”

Actively mentioned in recipe.
For example, a recipe for “grandma’s homemade oatmeal raisin cookies” shows intent for fresh-baked cookies in general (though the consumer has shown intent toward baking it themselves).

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Generic product reference.
For example, a recipe that specifically calls for butter, but does not specify a type or brand. 

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Explicit product recommendation.
For example, a professional (e.g. Dr. Oz), blogger (e.g. Recipe Girl), or influencer (e.g. @spoonforkbacon), writing an explicit recommendation about the product such as a post titled: "5 Reasons Why Butter is Best to Cook With."

Generic product coupon page.
For example, a Groupon page listing discounts for all types of butter.

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Specific product listing page.
For example, a search within a Retailer’s site for “high-fat butter”; or a category browse page (Dairy > Butter & Margarine > Butter).

Customer ratings/reviews page.
For example, a page with reviews for high-fat butter.

Specific product coupon page.
For example, a Groupon page listing coupons for high-fat butter.

Total count of URL views with this Intent score for the selected product.

This score is assigned when an Instance exhibits:

Product display page.
For example, a Walmart.com listing for [Specific Brand’s] Full-Fat Butter.

Data aggregated on the State level.

Data aggregated on the County level.

Expressed as YYYY-MM-DD.

12-digit UPC code.

Intent Delta measures the change in aggregate Intent from the previous time period to the selected time period. It can be expressed as a positive number, a negative number, or zero, depending on the scale and direction of a line item’s change:

• Positive scores (+) indicate an increase in Intent.

• Negative scores (-) indicate a decrease in Intent.

• Zero (0) — indicate no change in Intent.

For example, when viewing a report for the Butter Subcategory over the last month (e.g. June), its Intent Delta would be calculated by taking its Intent score over the selected month (June) and subtracting from it the Intent score from the month prior (May). If Butter’s Intent score increased from 380 in May to 400 in June, then it would have an Intent Delta of +20 for this report.

Intent (selected time period) - Intent (previous time period) = Intent Delta

Your organization’s actual Sales figures, in USD, integrated into our platform.

Your organization’s actual Sales figures, expressed in units sold, rather than USD.

Sales Delta ($) measures the change in total sales from the previous time period to the selected time period. Sales Delta ($) is expressed as a percentage, indicating the percentage change in Sales, rather than the change in raw Sales numbers.

For example, a line item whose Sales increased from $90 in the previous period (e.g. May) to $100 in the current period (e.g. June), would have a Sales Delta of 11.11%. This number indicates an 11.11% increase in Sales from the previous period.

This is calculated as the difference in Sales over the periods ($100 - $90 = $10), divided by the Sales for the previous period ($90).

10 / 90 = .1111 or 11.11%

When Sales decrease, Sales Delta ($) is expressed as a negative. For example, if June’s Sales totaled only $90, compared with May’s $100 Sales figure, the Sales Delta ($) would be calculated as follows:

90 - 100 = -10

-10 / 100 = -0.1 or -10%

This number indicates a 10% decrease in Sales over the selected date range.

Sales Delta (Units) measures the change in total sales from the previous time period to the selected time period, expressed as Units sold, rather than USD. It also differs from Sales Delta ($) in that it is expressed as the change in raw number of Units sold, rather than as a percentage change.

For example, if an item’s Sales increased from 1,500 Units to 2,000 Units from one month to the next, its Sales Delta (Units) would be calculated as:

Units sold (this time period) - Units sold (previous time period) = Sales Delta (Units)

2,000 Units - 1,500 Units = +500

Correlation measures the relationship between Sales and Intent for a particular line item.

• Positive Correlation (+) indicates that Sales and Intent move in the same direction—either Sales and Intent both increase:

OR Sales and Intent both decrease:

• Negative Correlation (-), on the other hand, indicates that Sales and Intent move in opposite directions—either Sales decrease while Intent increases OR Sales increases while Intent decreases:

Correlation will always be contained in the range of -1 to 1, with 1 indicating direct (positive) Correlation and -1 indicating inverse (negative) Correlation.

Line items with negative Correlation should be investigated.

Correlation Delta measures the change in aggregate Correlation from the previous time period to the selected time period. It can be expressed as a positive number, a negative number, or zero, depending on the scale and direction of a line item’s change:

• Positive scores (+) indicate an increase in Correlation.

• Negative scores (-) indicate a decrease in Correlation.

• Zero (0) — indicate no change in Correlation.

For example, a line item with a Correlation of +0.87 in the selected period (e.g. June), increasing from +0.83 in the previous period (e.g. May), would have a Correlation Delta score of +0.04.

Correlation (selected time period) - Correlation (previous time period) = Correlation Delta

Measures the relationship between Unit Sales and Intent. Positive correlation indicates that Unit Sales and Intent are moving in the same direction, either both increasing or both decreasing. Negative correlation indicates that Unit Sales and Intent are moving in opposite directions—either Unit Sales increases while Intent decreases, or vice versa.

Correlation will always be contained in the range of -1 to 1, with 1 indicating direct (positive) Correlation and -1 indicating inverse (negative) Correlation.

Line items with negative Correlation should be investigated.

Measures the change in Correlation (Units) from the previous time period to the selected time period. It can be expressed as a positive number, a negative number, or zero, depending on the scale and direction of an item’s change:

• Positive scores (+) indicate an increase in Correlation. 

• Negative scores (-) indicate a decrease in Correlation. 

• Zero (0) indicates no change in Correlation.

Correlation (Units) (selected time period) - Correlation (Units) (previous time period) = Correlation Delta (Units)

Visualization

Description

Bar

Use when comparing many aggregates/sums during a set time period.

Line

Use when comparing how several variables (<10) change over time.

Pie

Use when comparing several categories against each other.

Table

Use when a close analysis of several variables is needed.

Heatmap

Use when analyzing data geographically.

Number

Data expressed as numeral, rather than visualization.

Text/Markdown

Text/Markdown to help your data tell its story.

Image

Upload an image.