Anonymous Linchpin usage statistics

2. What is the purpose of Linchpin App Usage?

By means of (absolutely anonymous) usage data we, as developers, can learn how our software products are actually used. All such data serves as a supplement to regular user tests, project lessons learnt in the field as well as feedback from our customers.

This wide range of usage input allows us to more quickly decide which functionality is useful. Anonymous tracking helps us to make informed decisions regarding our future app development.

3. How does Linchpin App Usage work? 

3.1 Types of data

We distinguish between two kinds of collected data: "view" and "interaction".

3.1.1 View

Used to monitor certain events observed over a span of time, such as loading time or average staying time.

3.1.2 Interaction

A single, one-off action that is triggered e.g. by clicking a button or a link.

3.2 What kinds of data can be collected?

Only data conforming to a fixed JSON schema can be written to the database.

Anything else communicated or sent by the browser is neither stored nor saved or transmitted anywhere in any form.

Currently tracked events:

The schema is included below:

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "$id": "http://usage.seibert-media.io/event.json",
  "type": "object",
  "title": "The JSON Schema for validating events sent to usage.seibert-media.io",
  "required": ["metadata", "events", "apps"],
  "properties": {
    "metadata": {
      "type": "object",
      "required": ["sessionID", "environment"],
      "properties": {
        "sessionID": {
          "type": "string",
          "description": "A unique sessionID to link events"
        },
        "environment": {
          "type": "object",
          "required": ["browser", "resolution", "hostVersion", "maxUsers"],
          "properties": {
            "browser": {
              "type": "string"
            },
            "resolution": {
              "type": "string"
            },
            "hostVersion": {
              "type": "string"
            },
            "maxUsers": {
              "type": "integer"
            }
          }
        }
      }
    },
    "apps": {
      "type": "object",
      "additionalProperties": {
        "type": "object",
        "required": ["version"],
        "properties": {
          "sen": {
            "type": "string"
          },
          "version": {
            "type": "string"
          }
        }
      }
    },
    "events": {
      "type": "array",
      "items": {
        "type": "object",
        "required": ["kind", "name", "timestamp", "app"],
        "if": {
          "properties": { "kind": { "const": "view" } }
        },
        "then": {
          "required": ["kind", "name", "timestamp", "app", "loadDuration"]
        },
        "properties": {
          "kind": {
            "enum": ["view", "interaction"]
          },
          "name": {
            "type": "string"
          },
          "timestamp": {
            "type": "string"
          },
          "app": {
            "type": "string"
          },
          "loadDuration": {
            "type": "number"
          },
          "data": {
            "type": "object",
            "required": [],
            "properties": {},
            "additionalProperties": true
          }
        }
      }
    }
  }
}

3.3 How would an example set of data look like?

The table below shows three complete records, containing all information that is written to the database.

How is this data evaluated by Seibert Media?

A sample "interaction" event processed with Google Data Studio

3.4 How frequently is data stored and transmitted?

The collected information is kept in the session storage and assigned a random session ID. This ID is assigned per browser tab. Sessions can never be associated with more than the session ID itself, i.e. they never contain personal data.
A data packet is sent to the database whenever five (interaction or view) events have been registered. Upon leaving a page with events, the remaining data is sent.

Sent data is not visible in the cache or in sessions after being transmitted ("fire and forget"). The data is sent to a proxy maintained by Seibert Media. It validates the transmitted information according to the schema described above, and thus ensures that we can only gather data in accordance with that schema.

All information validated by our proxy is written to the database, which is hosted on a Google Cloud server.