A useful freelance portfolio is evidence of a repeatable process, not a collection of attractive claims. Beginners can demonstrate Excel and reporting skills with clearly labelled synthetic data, provided they never present the exercise as a paid customer project. This guide combines official Fiverr portfolio rules with Microsoft's learning resources and a practical acceptance-test workflow.
1. Start with a measurable buyer problem
A first portfolio does not require a past client, but it does need truthful evidence of a specific skill. Choose a small task someone might actually purchase: cleaning repeated sales rows, preparing a basic expense dashboard, or documenting an Excel automation. State what was wrong with the initial data, what the output should achieve, and how you will test it. Do not combine ten unrelated services into one showcase. For this tutorial, imagine a fictional shop invoice table created from scratch, without real customer names, account numbers or business transactions. The point is to demonstrate a process, not to simulate a prestigious client relationship.
Write acceptance criteria before opening your spreadsheet: duplicate invoices should be identified, totals should reconcile after documented exclusions, and a new month should be refreshable without rebuilding every formula. These are more informative than calling a dashboard 'professional'. Label every image and deliverable 'synthetic demonstration data' where a viewer could mistake it for real work. Never attach a real company logo or invented testimonial. A simple, reproducible example often proves more than a visually impressive dashboard whose calculations cannot be explained.
2. Create synthetic data without privacy shortcuts
Build an invented dataset with invoice ID, date, category, quantity, unit price and payment status. Use fictional identifiers such as INV-001. Do not copy a confidential customer database and merely replace the names: detailed dates, unusual amounts and locations may still identify people. Create observations from scratch or use a learning dataset only after checking its reuse conditions. Microsoft's published Financial Sample workbook is useful for learning Power BI, but a public download does not automatically grant permission to resell the workbook as your own product. Distinguish learning from redistribution and describe the data provenance accurately.
Check the sample before charting it. Include a few deliberately wrong entries in a 'before' version, such as a duplicate invoice, an invalid date and an inconsistent amount. Explain the correction rule and keep both the raw and cleaned files. If you remove three duplicates from a synthetic table, you may state that measurable result for the demonstration. Do not transform it into a fabricated claim that you saved an actual company 40 percent of its costs. Keeping the example small makes it easier for a buyer to inspect the transformation and understand your decisions.
3. Make the demonstration independently testable
Package three elements: a small data file, a finished Excel or Power BI result, and a short readme describing refresh steps. Show one meaningful indicator rather than a crowded wall of unrelated charts. Explain important formulas or Power Query transformations and describe what happens when a required column is missing. Document the assumptions behind currencies, returns, date ranges and categories. If the solution only reads a local training table, do not claim integration with a live accounting system. The most persuasive evidence is a clear relationship between the buyer's problem, the method and a result that can be reproduced.
Create an acceptance test: add one new invoice with a known amount, refresh the report and verify the resulting total against the original sum. Fix discrepancies before taking screenshots. Keep before-and-after views at the same scale so the improvement is not a trick of presentation. Remove sensitive metadata from files and screenshots. Attractive visuals are not proof of correct calculations, which is why a brief test note and an example of an error case add more credibility than another decorative chart.
4. Use Fiverr's portfolio without inventing clients
Fiverr's official help documentation explains that sellers may add portfolio projects manually, including work completed outside Fiverr, and that projects can appear on their profile and Gig pages. That does not make it acceptable to describe a training exercise as a paid customer order. Title the sample 'Synthetic invoice-cleaning demonstration' and describe the problem, tools, deliverables and limitations. Do not attach a fake review, invent a customer name or imply that a recognisable brand hired you. Fiverr's community standards also prohibit presenting another person's work as your own and require client consent when reusing work from a real order.
Use two or three complementary images: one shows the result, another shows the transformation, and a third explains what is included or excluded. A generic laptop photo cannot substitute for a screenshot of the actual sample. The licensed photograph accompanying this article illustrates a workspace; it is not evidence of work performed by RibhLab. Keep portfolio descriptions factual and avoid promises of ranking, sales or a particular number of inquiries. Review the platform's current requirements before uploading because interface options and presentation limits can change.
5. Protect files, licenses and confidential information
Inspect file properties for computer usernames, private folder paths, hidden sheets and internal comments. Remove external database connections, API keys, passwords and links to customer systems. Check the license of every third-party icon, photograph, font and template. Avoid using brand marks in ways that imply an endorsement. If you want to show a real client's work, obtain explicit permission about the material and level of detail that may be published. Otherwise build a new example with independently generated data rather than transferring confidential records into a 'sanitised' presentation.
Where possible, retain both an editable source and an accessible preview that does not require paid software. State technical requirements, known limitations and the absence of live data connections. Do not distribute unnecessary macros or executable components. Record the creation date, revision date, original components and licensed external assets. This small provenance log makes it easier to answer ownership questions and prevents a demonstration from being mistaken for a completed commercial engagement. If reuse rights are uncertain, leave the questionable material out until they are established.
6. A seven-day production and review plan
On day one, select one service and its acceptance criteria. Use days two and three to create and clean the fictional dataset, and day four to build the report. Reserve day five for independent tests, day six for images, descriptions and license checks, and day seven for a buyer-perspective review. Ask someone unfamiliar with the work to explain its purpose after a minute of viewing. If they cannot, simplify the presentation. No purchased reviews, invented customers or unrealistic income claims are needed. The first success is a demonstrable skill sample, not a guaranteed order within a week.
After publishing, track two separate questions: does the sample work correctly, and can prospective buyers understand the offer? A technically excellent example can still have an unclear scope or price, while a polished image may hide a weak process. Update the description, buyer questions and delivery requirements using real feedback. Change one variable at a time when experimenting, and avoid drawing strong conclusions from a single view or inquiry. A credible portfolio shows what you can do now and openly states what you have not done.
Official resources
- Fiverr Help Center — Using your Fiverr Portfolio (checked 30 October 2026) ↗
- Fiverr Help Center — Community Standards, original work and consent (checked 30 October 2026) ↗
- Microsoft Learn — Financial Sample Excel workbook for Power BI (updated 9 September 2025) ↗
- Unsplash — image license and commercial use ↗
No online income or trading system guarantees profits. Results depend on skills, costs and circumstances.
Disclaimer ↗