Type A: Power user, data analyst

A power user (often the data analyst) typically has three stages of usage of any analytics tool:

1. Retrieving and preparing data

• Pull data from various sources: Azure, Redshift, Snowflake, PostgreSQL, GA, MP, Segment, HRMS, Twitter, Nielsen tools, local files, etc.

• Prepare data: Data permissions (users, groups, org.), harmonisation, joins, aggregation, security, etc.

• Semantic layer (only usable if an advanced query system is to be made available for business users): Synonyms, descriptions, relationships, etc.

2. Putting together insights (and supporting charts)

• Pre-defined set of KPIs, trends or views of charts that needs putting together: dashboard tool or Excel.

• Supporting data powering these charts or KPIs are also included in shared excel reports. This is also provided to do additional checks and validations.

• Power BI, Qlik, Tableau are most popular pre-configured ways of putting together organisation business information. MP, Segment, GA are other tools that are used heavily to put together product analytics information together; more on this later in Type B of users.

• These tools also have started providing NLG based information retrieval systems for sentence-format answers and speed up the process for report or dashboard creation.

3. Present to or share with stakeholders (business experts)

• The work-in-progress dashboard is then shared with functional or business owners for review.

• Connected dashboards in the form of tabs are built within the primary dashboard. This is based on requirement.

• Once reviews with change requests are fulfilled, the dashboard is handed off. In most cases, primary consumers are added as Viewers, Analysts and creators are admins, editors.

Type B: Functional owner (Head of Sales, Head of Marketing, etc.)

1. Delegate + collate data:

• Pull data from various dashboards (not direct sources). This is often either delegated to junior analysts or self, depending on who the tool owner is. The primary purpose here is to collate information to start putting together a report.

• Time, effort, attention is usually derived from an end-goal: Meeting. Type of meeting and distance (time) to meeting. E.g. Quarterly NPS heartbeat, Weekly Product Catchup, etc.

2. Put together charts & insights:

• End deliverable is mostly observed to be either Excel, PPT, or PDF.

• Dashboards are not typical in these meetings since the mental model of consuming these are focused and objective-oriented. While dashboards serve the purpose of 'regular' monitoring, Excel, PPTs, or PDFs have a deeper theme associated with them.

• In most cases, insights and charts are presented as screenshots with emphasis on period, comparison (growth, distribution), outliers, summaries, and trends.

3.Present or Share with stakeholders:

• The work-in-progress dashboard is then shared using email primarily.

• At the time of presenting - they're used as references of facts and insights - nothing more.

• They're also used occasionally as anchor references when further deep-dive is required; this is when a more focused dashboard is fired up to answer questions.

Type C: Business user, consumer, decision maker

1. Consume data:

• Consumed in emails and on mobile primarily. Have also seen screenshots being shared in the IM apps like Whatsapp/ Slack in focused groups.

• A very popular email format is long form style analysis reports. The report contains screenshots of charts and usually accompanied with very contextual executive summaries. Any accompanying assets, references are attached as part of email attachment.

• Attachments include PDFs, PPTs, Excel, Images.

• The process of analysis is typically not part of the original shared content. This is typically added when asked for. Analysis path for a pre-determined, frequently viewed artefacts also are commoditised over time.

• The medium of data consumption is driven by where the primary decision maker, consumer is on: Email, Slack, Whatsapp, etc.

• Consumption is not straight forward here. It's simply not 'view-only'. Consuming data invokes further clarifications, questions which is where the next step comes in.

2. Request deep-dive, collateral information:

• Double-clicking on a KPI for further details is a standard. What's the bird-eye view? Why's a number green, red, or constant? And many more as such.

• Some other contextual data are also required which were not possible to account for earlier. E.g. campaign imagery for seeing drivers and impact in a trend, app screenshots for references, etc.

3. Decide and move on:

• Instruct a set of next steps to work on - typically different business holders take it from there. E.g. A global CPG marketing may follow the following path: Business dashboards → Questions, analysts → Reports → Alignment → Product SKU owners → Creatives for next campaign

• The initially shared (and evolved with threads + to-fro) is only referred back once newer improvements are made.

• Stakeholders are constantly moving from one piece that needs decision to the next.