The deliverables that turn a pilot into a lasting foundation

IN BRIEF  A good starter pack gives you a report, shared definitions, a model people can understand, a repeatable data flow, documentation and people who can keep it running.

A data starter pack should help you learn quickly on a real use case beyond the demo. We expect a good pack to clarify your indicators, work with the sources you have today and give you a first dashboard your team can judge on the facts. The handover matters as much as the deliverables.

When comparing offers, look at what you will have at the end of the pilot. Will you have validated definitions, a readable model, a repeatable refresh, documentation and someone trained to use them? If every change requires the provider, you may have received a report with little to build on.

A “pack” can be a simple demo, a bundle of licences or an end-to-end service. Ask what you get to keep, which data is actually connected, what runs automatically and what your team can do independently.

The starter pack should pair a visible result with the foundations it needs to keep working.

1. An initial assessment and a priority use case

The initial assessment should help you make a decision. We want to understand the process, the sources and the problem before promising a dashboard. Sometimes the best outcome is to narrow the scope or fix a data entry step upstream before building anything.

The assessment checks whether the pack suits your systems, baseline data quality, user availability, constraints and goals. It may conclude that the format is unsuitable. The chosen use case describes a decision, a process, a group of users and a measurable outcome.

2. A set of three to five KPIs

The pack should define the indicators as well as displaying them. For each KPI, specify its formula, scope, source, frequency, target and owner. Three to five indicators are enough to test value while keeping the pilot manageable.

3. A map of your data sources

The map documents systems of record, files, keys, owners, update frequency, history and data quality limits. It reveals dependencies and missing data. Even a brief map should let another person trace every figure to its source.

4. A basic data model

The model centralises the rules and separates facts, dimensions and measures. It should be documented, use business terminology and allow for future changes. A dashboard connected directly to a file can work in a demo, but it gives you little to build on.

Scoping, KPIs, the model, the dashboard, governance, training and a roadmap belong together.

5. A working dashboard with real data

Dummy data can help you discuss the design, but it cannot establish the quality of the sources or the accuracy of the rules. A starter pack should test the model early with a familiar month, a specific exception and a future user. That is how you build trust.

The visible deliverable is a report the people in charge can use. It includes a summary, simple filters, a trend and a detailed view that explains deviations. It should show the refresh date and be tested with users.

6. A first repeatable data flow

The pack should state how the data is refreshed. Some offers include an initial manual load and sell automatic refreshes as an option. In all cases, ask for clear instructions on where to find the data, the format, the schedule, the checks, what to do if a run fails and whom to notify.

7. Security and governance rules

Even a pilot should identify sensitive data, authorised users, the model owner and retention rules. Check sharing and export permissions. A few clear rules now spare you the work of finding out later who owns each report and which version to trust.

8. Training and handover

We want you to be able to explain where a figure comes from, who knows the source and what happens if a refresh fails. Training should cover how to use the dashboard, how it works and where its limits are. Independence is part of the result.

Training shows users how to read, filter, verify and use the dashboard in an actual decision. The handover includes documentation on the sources, KPIs, refresh process and incident contacts. A demo video alone cannot do that work.

9. A roadmap after the pilot

The roadmap identifies necessary fixes, worthwhile extensions and optional investments. It might propose new sources, automation, hosting, security or training. Rank items by value and effort before committing to them.

A checklist for comparing data starter packs

ComponentExpected evidenceQuestion to ask
ScopeUse case and success criteriaWhat decision will improve?
DataMap of sources, keys and limitationsWhich source is authoritative?
ModelDocumented tables, relationships and measuresCan the model be reused?
DashboardReport tested with real dataWho validated the figures?
RefreshInstructions, checks and alertsWhat happens if a run fails?
HandoverTraining and documentationWho takes over after the engagement?
Next stepsPrioritised roadmapWhat can safely wait?

What Datavizin lists in its Data Starter Pack

According to the page consulted in July 2026, the Industrial Data Starter Pack is intended for manufacturing SMEs with 20 to 250 employees and offers a dashboard in 10 working days. It includes a free one-hour initial assessment, audit and scoping, data modelling and centralisation, a Power BI dashboard, two hours of training and handover, and recommendations for what comes next.

The page lists the pack at CHF 1,750 and an optional CHF 149 monthly subscription for automated data flows, sharing, sovereign Swiss cloud hosting and infrastructure management. Deliverables, terms and prices may change. Confirm them on the official Data Starter Pack page before publishing or making a decision.

What a good pack spells out about exclusions

A clear scope protects both sides. Make sure you know whether extensive historical data cleaning, a specific connector or fine-grained security requirements fall outside the pack before work starts. Clearly stated limits are easier to manage than exceptions to a vague promise.

Check whether the pack includes a full migration of every source and historical dataset, comprehensive enterprise data governance, real-time operation when sources cannot support it, KPIs with unconfirmed definitions, full automation without the necessary infrastructure, licences or access, or a commitment to outcomes based on inaccessible or unusable data.

How we measure the success of the pack with you

The dashboard answers the priority question with validated figures. Preparation or search time goes down. Users can filter, verify and act. Sources, KPIs and owners are documented. There is a clear procedure for refreshes and incidents. You can estimate the budget and effort for the next extension before deciding whether to start it.

The ECSTAA model gives you a framework for exploring, understanding, structuring, analysing and acting. Datavizin’s plans then let you choose infrastructure or a subscription suited to your needs.

NEXT STEP  Compare offers using the checklist above. Ask to see a sample deliverable, the refresh procedure and who will be responsible for what after launch.