Commerce Reporting Glossary

Unlike reports that provide an overview of the current performance or performanace development of a company, a company department or certain business activities, analyses provide answers on concrete business questions and help develop actions to solve problems or maximize performance. A cohort analysis for example breaks down customers and their purchasing metrics into cohorts and compares them with each other – like that, it answers questions on potential patterns within customers’ buying behaviour and helps to develop actions that maximize performance.

Attributes are additional categories of information assigned to the entities within a data model. Age, gender and order country, for example, are attributes at the entity “customer”. Similar to like they should with entities (although in a limited extent), companies should also be able to conduct certain analyses on single attributes resp. attribute values (for the examples mentioned above, e.g. 25, male, France) – that is to say: break down certain metrics on single attributes resp. attribute values.

Compilation of various visualization modules like charts, graphs or diagrams into one comprehensive one-page overview. A dashboard is often composed from single tiles or widgets with one visualization module on each of them.

Holistic systematization of metrics, entities and attributes that works as a meta model for structured data base like e.g. a data warehouse. Analytics data models in particular define the calculation of metrics, the availability of entities and attributes as well as the logical connection between all those values and dimensions. Thus, it ultimately is the data model that determines which data can be accessed and used in which way and in which level of detail.

Databases of operational systems (like e.g. eCommerce platform, order management or campaign management system) that push their data (or have it pulled) into some kind of data lake, data processing architecture or other operational systems to make it usable for analytics and/or automation purposes.

In our model, markdowns are defined as all kinds of changes in price that are implemented globally. A discount is a change in price that is not valid globally, but for single customers (or customer groups) only. Examples: If a retailer decides to generally sell a product 10€ cheaper than before, that’s a markdown. If a retailer implements individual price calculations or directly promotes a product to a certain customer group for a price that is different from the general price (e.g. via vouchers or coupons), that’s a discount. In cases of general changes in price being valid during certain time spans (e.g. “this week” or “on Mondays”), we recommend declaration as discount, as those are promotion offers that are only valid for the group of customers who buy during that certain time span – but ultimately, it depends on how you want to separate both types of reductions in your reporting. The most important thing is: Be consequent.

Central, structured database that is optimized for analytics and/or automation purposes, e.g. by allowing comprehensive aggregation across multiple dimensions as well as data access in a very fine-granular level of detail. In most cases, a DWH is filled from multiple, heterogeneous data sources and, in analytics scenarios, often feeds some kind of analytics or data visualization frontend that enables users to gain insights from their data in a fast and easy way.

In a non-technical, but analytical sense, an entity is an object that, data-wise, can be looked at in isolation. In the context of commerce reporting, key entities are e.g. customers, products, orders or campaigns, so companies should be able to report on each of those – that is to say: drill-down their metrics to their single manifestations (e.g. display performance metrics of single products or single campaigns). An elaborate data model is key to ensure that all important entities can be accessed and analyzed in the required way.

Regarding the differentiation against the term “attribute”, see the according glossary entry.

Our model considers all costs as fees that a customer pays to realize a transaction. See our transaction metrics matrix for classification of fees within our transaction metrics model.

Within the Commerce Reporting Standad, the terms “gross” and “net” do not refer to before or after taxes, but to before or after returns – as this is the relevant differentiation for business cases in commerce companies. Regularly, we spare the “gross” in our metric terminology and only explicitly use “net” to indicate “after returns”. See our transaction metrics matrix for classification of gross and net metrics within our transaction metrics model.

A KPI (Key Performance Indicator) is a value from the total quantity of company-relevant metrics that is critical for business performance evaluation and, as such, is often looked at in relation to defined target figures. In direct comparison with those target figures, the actual KPI values can be evaluated and, if necessary, developed in the desired direction with the help of business activities derived from the data.

In addition to basic KPIs like revenue or contribution margin, many commerce companies heavily rely on additional metrics like e.g. customer lifetime value and include those into their set of KPIs – in eCommerce, there are some purely online-relevant KPIs on top of that, e.g. cart abandon rate. Furthermore, there are metrics that are typically used as KPIs by certain commerce industries, but not by others – return rate is a good example for that: This metric does often have KPI status in fashion companies, whereas for food or furniture companies, it’s mostly treated as one of many metrics in weekly and monthly reporting. Moreover, KPIs are also defined department-specific: While e.g. customer lifetime value is critical for CRM, click-through rate is likely to only have KPI status for marketing staff. Ideally, at least parts of the company KPIs are handled completely company-specific to make sure they match the company goals in the best possible way.

Attention: Often times, the term “KPI” is mixed up with the term “metric” – the terms are not the same, though. KPIs do only represent one sub-segment of the total quantity of metrics and are at least industry- and department-specific, in many cases they’re even defined individually per company. See also “Metric”.

In our model, markdowns are defined as all kinds of changes in price that are implemented globally. A discount is a change in price that is not valid globally, but for single customers (or customer groups) only. Examples: If a retailer decides to generally sell a product 10€ cheaper than before, that’s a markdown. If a retailer implements individual price calculations or directly promotes a product to a certain customer group for a price that is different from the general price (e.g. via vouchers or coupons), that’s a discount. In cases of general changes in price being valid during certain time spans (e.g. “this week” or “on Mondays”), we recommend declaration as discount, as those are promotion offers that are only valid for the group of customers who buy during that certain time span – but ultimately, it depends on how you want to separate both types of reductions in your reporting. The most important thing is: Be consequent.

A metric is a directly measurable value or a value composed from multiple directly measurable values that is used for performance measurement of single business activities, business components or entire company departments. For the Commerce Reporting Standard, wie differentiate between six blocks of metrics: transaction metrics, customer metrics, product metrics, operations metrics, journey/acquisition metrics and on-site/funnel metrics.

Attention: Often times, the term “metric” is mixed up with the term “KPI” – the terms are not the same, though. KPIs do only represent one sub-segment of the total quantity of metrics and are at least industry- and department-specific, in many cases they’re even defined individually per company. See also “KPI (Key Performance Indicator)”.

The term “raw data level” is often used in the context of data accessibility especially for operational roles, because while on higher levels of the company hierarchy, aggregated data in the form of overview reports or single KPIs is of particular interest, e.g. a CRM Manager does also need his or her data broken down to single customers – that is to say: on raw data level. Thus, the term names the level of detail in which data sets are not (only) available in an aggregated way, but also, kind of, in their pure form.

Unlike analyses that provide answers on concrete business questions and help with the development of business activities to solve problems or maximize performance, reports provide an overview on the current performance or performance development of a company, a company department or certain business activities. Often times, reports are all about presenting the performance of a certain department to the superior level (e.g. an eCommerce Director’s monthly business figure report to the corporation or the Head of Marketing’s weekly KPI overview to the CMO), but also employees with purely operational work become report recipients, when they find the automatically generated daily report from their day-to-day tool in their inbox every morning (e.g. the email marketing system’s report on open and click metrics of the previous day’s campaigns).

Whereas the term “report” refers to a certain document or data view, the term “reporting” usually refers to the general activity that is behind producing a report – e.g. the daily, weekly or monthly routine of providing a report to the superior level. Also, the term can be used for a certain collection of reports, metrics and/or KPIs like e.g. the “monthly marketing reporting” or the “daily sales reporting”.

Every transaction-related metric needs a defined time reference as there is not only one, but multiple dates assigned with an order – first of all: the order date, the invoice date and the return date. Depending on the selected time reference, the set of orders that is accounted for the chosen time span changes. Click here to read the detailed explanations in our forum.

Visual display of data in easily understandable formats such as reports, charts, graphs or diagrams. Single visualizations are often composed to dashboards.