Qlik Alternative: Priced by the Gigabyte, Steered to the Cloud

You Are Now Billed for Gigabytes, Not People

In March 2025 Qlik completed a move to capacity-based pricing, and it is worth understanding exactly what that means because it is unusual.

Qlik Cloud Analytics lists at $300 a month for Starter, $825 for Standard, and $2,750 for Premium, billed annually. Only Starter counts users, and it caps them at ten. Every tier above it advertises unlimited users and charges instead for data for analysis — 25 GB on Standard, 50 GB on Premium, topped up in 25 GB or 250 GB packs.

Read that again, because the consequence is easy to miss. Your bill is no longer driven by how many colleagues log in. It is driven by how much of your own data you are allowed to look at. Adding two years of history to a model is now a purchasing decision. So is keeping a wider fact table. So is onboarding a large customer.

Qlik's own product is built on loading data into memory and letting people explore it freely. Metering that by the gigabyte puts the pricing model in direct tension with the way the software is meant to be used.

And whatever you think of the model, note that it replaced the one you originally bought. The terms moved.

Self-Hosting Is Still Possible. It Is Just Not on the Menu.

To be straight about it: you can still run Qlik on your own infrastructure. Qlik Sense Enterprise on Windows exists and is supported.

What has changed is its standing. Every published price on Qlik's pricing page is a Qlik Cloud price. Self-hosting no longer appears as a listed plan at all — the page ends with a single line inviting you to contact sales if you need to host on your own IT infrastructure.

That is a clear signal of where a vendor wants you. Cloud is the default, the documented option, the one with a number next to it. On-premises has become the exception you have to ask for, which in practice means a negotiation rather than a price.

If you are costing a new Qlik deployment today, you are costing Qlik Cloud whether you intended to or not.

QlikView Keeps Losing Pieces

Be accurate here, because plenty of pages on the internet are not. Qlik has explicitly said it has no plans to end-of-life QlikView and no plans to force anyone onto Qlik Sense. QlikView 12.100 shipped in September 2025 with support running to September 2027. Anyone telling you QlikView is dead is guessing.

What is documented is narrower and more useful. QlikView Source Control and QlikView Workbench stop being supported or updated from the September 2026 release. Source Control is how QlikView developers connect the desktop to version control. Workbench is how QlikView integrations get built in Visual Studio. Neither is exotic — they are the developer tooling around the product.

Existing releases keep working. But a product that keeps the core and quietly retires the tooling around it is telling you where it sits in the roadmap. QlikView is being kept, not advanced, and the investment plainly runs to Qlik Sense and Qlik Cloud.

And It Is Expensive at Both Ends

At small scale Qlik is not merely pricey, it is the expensive option by a distance — $300 a month buys ten users and ten gigabytes. At large scale the complaint inverts and becomes about growth: teams describe it as too expensive to extend to hundreds or thousands of people, and implementation complexity generates costs that were never in the original quote.

What DataPallas Does Instead

It is one piece of software that explores your data, builds reports from it, delivers them, runs a portal for the people receiving them, and serves dashboards and embedded analytics. Open source, self-hosted on hardware you own, reaching any JDBC database, at home on Linux, Windows, or macOS.

DataPallas is open source with fair commercial pricing, and what it is priced around is the people who build and administer the reports — not the people who read them, and not the volume of data you load. Readers are not a licensed tier and gigabytes are not metered, so putting a dashboard in front of a thousand people is a distribution decision rather than a purchasing one.

Your Data Does Not Go Anywhere

Queries are answered by your own database, from a machine inside your own network. There is no volume to declare, no extract to schedule, and no ceiling on how much history you can put in front of someone — which removes the strangest part of capacity pricing, where the cost of an analysis depends on the size of the question.

For anyone working under GDPR, HIPAA, financial regulation, or data-residency rules, keeping the data in place also makes the security review considerably shorter.

The Associative Bit, Handled by the Model

Qlik's strength is exploration: click anything, everything else responds. The equivalent structure here lives in Cubes — dimensions, measures, joins, segments, and hierarchies declared once, so relationships are described in one place rather than rediscovered per app. Reports, dashboards, AI questions, and embedded components all read it, and AI drafts it from your live schema, so you are not starting from an empty editor.

Source Control That Cannot Be Withdrawn

Qlik is retiring QlikView Source Control, which is a fair reminder of how awkward versioning has always been around .qvw and .qvf: they are binaries, so there is nothing to diff and nothing to merge.

Here there is no source-control add-on to lose, because none is needed. You configure everything through a UI — source, template, parameters, schedule — with optional AI help beside the fields that want code. What that UI saves is plain text: the cube definitions, the queries, the templates. They live in Git with the rest of your code, get reviewed before they ship, and roll back cleanly. The interface stays easy and the history stays intact.

The Reporting Half Qlik Sells Separately

NPrinting exists because Qlik Sense does not produce and distribute documents on its own. Here that is not a second product.

Reports come from SQL, Groovy, or files, and emerge as PDF, XLSX, DOCX, HTML, or any text format. Layouts are written in HTML and CSS — real page breaks, repeating headers and footers, exact placement, web fonts, and a preview that redraws when you save — with XSL-FO for demanding typography.

Then bursting: one report split by customer, employee, branch, or cost centre and delivered over email, FTP/SFTP, cloud storage, HTTP/WebDAV, or a portal, on a timetable, with retries and checks before anything is sent. The Document Portal gives every recipient their own documents and nobody else's.

Dashboards, Embedding, and Scale

Live dashboards, plus data tables, charts, and pivot tables that drop into your own application behind one script tag.

When analytical load would hurt production, CDC replication feeds DuckDB or ClickHouse on your own hardware — the speed of an in-memory model without a gigabyte allowance attached. Chat2DB turns questions into SQL, and the AI crew helps across the rest.

The Honest Bit

Qlik's associative engine is genuinely distinctive. Selecting a value and instantly seeing which related values remain possible — and which have been excluded — is something most BI tools do not really replicate, and analysts who think in that model are fast in it. If free-form associative exploration is what your team does all day, that is Qlik doing the thing it was built for.

Nothing here imports a .qvw or a .qvf, and load scripts are not translated. Applications are rebuilt, with the data model becoming a cube and the load script becoming SQL or a CDC feed. Real work — scope it first.

Phase One: Take Over Distribution, Keep Qlik Running

If you use NPrinting, or have built scripting around getting Qlik output to people on a schedule, that is the cheapest thing to move and the easiest to prove.

DataPallas can split and deliver documents from any source. Qlik carries on producing exactly as it does now; DataPallas takes over dividing the output by recipient, scheduling it, delivering it, and hosting the portal.

One production workload moved inside a week, no application rebuilt, nothing to unwind — and a working look at the platform against your own data before anyone commits further.

Phase two is the applications themselves, one at a time, on your schedule.

Where to Start

1. Find out which tier you are on and how many gigabytes you are paying for. Then work out what a year of extra history would add.
2. Check whether anything you run depends on QlikView Source Control or Workbench. Support for both stops at the September 2026 release.
3. Move one NPrinting-style distribution job to DataPallas, with nothing upstream changed.
4. Rebuild one small Qlik application — data model as a cube, load script as SQL — and compare the numbers.
5. Then weigh a capacity renewal against a platform where neither your users nor your data are metered.