Analytics & BI Tools · Chapter 16 of 18
The wider landscape: who sells what, and what changed
Sixteen products beyond the big three, the six renames that make older material wrong, and what it tells you that half the market stopped publishing prices.
Most BI writing covers three products. The market is much wider, and several of the interesting positions are held by tools nobody teaches. This chapter is a map, and it is deliberately organised around pricing model rather than features, because the pricing model determines which organisations a tool suits far more reliably than the feature list does.
Figures in this chapter were verified on vendor pages in August 2026 and they change constantly. Treat them as the shape and the ratios rather than as a quote, and check before you put a number in a business case.
Where a vendor does not publish a price, this chapter says so — because ‘not published’ is itself a finding, and the pattern of who publishes and who does not is one of the more revealing things about this market.
The four pricing models
The volume leaders, with current numbers
| Product | Price | The thing to know |
|---|---|---|
| Power BI Pro | $14 user/month, annual | Rose from $10 in April 2025 — the first increase since launch. Any material quoting $10 predates it |
| Power BI Premium Per User | $24 user/month | Does not provision a Microsoft Fabric capacity — a common and expensive misunderstanding |
| Fabric capacity | ~$0.18 per capacity-unit-hour, or ~$938 per CU-year reserved | Reserved is exactly 40.5% off. The 3-year reservation is the same annual rate as the 1-year — unlike Azure compute, longer lock-in buys nothing |
| Tableau Creator | $75 (Standard) / $115 (Enterprise) user/month | Roughly 5× Power BI Pro at the authoring seat |
| Tableau Explorer | $42 / $70 | The tier organisations underestimate — a 'viewer' who wants their own view needs this |
| Tableau Viewer | $15 / $35 | The Enterprise viewer more than doubles. At a large read-only audience this line dominates the bill |
Above Tableau's published tiers sit two unpublished ones bundling agentic features, and above Power BI's sit the Fabric capacities. In both cases the published ladder stops exactly where the interesting money starts, which is the pattern throughout this market.
The specialists, and the position each actually holds
| Product | Price | Its actual position |
|---|---|---|
| Qlik | $300 / $825 / $2,750 per month by tier | Moved off per-seat entirely. Priced on data capacity with unlimited users above the entry tier. Any material teaching 'Professional vs Analyzer seats' describes a retired model |
| ThoughtSpot | $25 / $50 user/month | Search-first. Also caps on row count, not just seats — and states plainly that it does not meter or charge for LLM tokens |
| Looker | Withdrawn — all three editions now 'call sales' | The model-first tool. Still the strongest governance story, and you can no longer price it without a salesperson |
| Strategy (was MicroStrategy) | from $13 user/month | The cheapest published enterprise seat here, and a deliberate repositioning. Note the parent company's main business is now elsewhere — relevant to a vendor-viability conversation |
| Amazon Quick Sight | Author $24, Reader $3, or $250/mo for 500 reader sessions | The only genuine pay-per-session reader pricing. Structurally cheap for large infrequent audiences and expensive for daily drivers. Watch the $250/month infrastructure fee that makes small deployments uneconomic |
| Metabase | Free self-hosted; $100 / $575 per month cloud tiers | Genuinely good open-source option. The catch: embedded end-users count toward your user total, which is what makes embedding expensive |
| Preset / Superset | Free tier; $20 user/month | Apache-licensed engine with a managed service. The enterprise AI feature requires you to bring your own model API key — an unusual and honest licensing choice |
| Hex | $36 / $75 per editor | Notebook-first. Prices editors only, viewers are free, and it publishes its compute profile rates — rare transparency |
| Sigma, Domo, Omni, Mode | Not published | Domo's model is consumption credits with unlimited users, which is genuinely differentiated. Mode is now part of ThoughtSpot |
What it means that half the market stopped publishing
In 2022 most of these vendors published a price list. Today Looker has withdrawn all three editions, Sigma redirects its pricing page to a contact form, and Alteryx publishes only its deliberately-limited entry tier.
The vendors that publish list prices are the ones confident of winning a price comparison. Microsoft publishes everything, including a public pricing API you can query programmatically. Strategy publishes a $13 seat precisely because it is competing on price.
The practical consequence for a buyer: you cannot build a comparison spreadsheet from public information any more, and the sales process is now the only way to get half the numbers. Budget time for that, and get quotes from at least two vendors in parallel — the discount available is a function of whether they believe you have an alternative.
And be sceptical of third-party pricing pages. Broker and review sites publish figures for vendors who publish nothing, and those figures are guesses or a single leaked contract. If the vendor does not say it, treat it as unverified.
Six renames that make older material wrong
| Was | Is now | Note |
|---|---|---|
| Looker Studio | Data Studio | A 2026 reversal of a 2022 rename, to separate the free personal tool from governed Looker. Any course saying 'Data Studio was renamed Looker Studio' is now backwards |
| Amazon QuickSight | Amazon Quick Sight, inside a broader suite | Capabilities, access and permissions unchanged — but every URL and doc reference moved |
| MicroStrategy | Strategy | Trading name in Feb 2025, legal name in Aug 2025 |
| Power BI 'datasets' | semantic models | Renamed in late 2023 and still the most common stale term in teaching material |
| Fabric 'license mode' | workspace type | Terminology only; functionality unchanged |
| Salesforce Data Cloud | Data 360 | Relevant because Tableau's newer agentic product builds on it |
The three structural changes worth understanding
1 · The capacity SKUs are consolidating. Microsoft has stopped selling new Power BI Premium per-capacity subscriptions; existing agreements run to term and the successor is the unified Fabric capacity. Per-user Pro, Premium Per User and the embedded SKUs are unaffected. If you are planning a renewal, the migration path is the conversation, and the post-expiry behaviour is worth knowing: a grace period, then throttling, then rejection of all operations, with data retained but inaccessible.
2 · AI is being metered, and the vendors have split. This is new and it is a genuine axis of differentiation rather than marketing.
- Looker meters conversational analytics by data token, at roughly $3 per million input and $20 per million output tokens, with generous included quotas and enforcement beginning 1 October 2026. That is the first hard date at which a major BI vendor puts a token line item on the bill.
- Databricks moved its conversational analytics to pay-as-you-go with a free monthly allowance per user.
- ThoughtSpot states explicitly that it does not meter or charge for LLM tokens.
- Microsoft went the other way on access, dropping the minimum capacity for its assistant from the F64 tier to F2 in 2025 — a roughly thirty-fold reduction in the entry cost of AI features, which changed the economics completely and makes any pre-2025 guidance on this badly wrong.
- The buying implication: if your organisation will use natural-language analytics at volume, the token policy is a first-order cost question and it is not on any comparison spreadsheet you will be shown.
3 · The preparation layer was repackaged. Alteryx moved from buying Designer seats plus Server cores to a platform model: an edition, then user roles, then a purchased capacity of automation runs. The entry tier is cloud-only and limited to flat files, deliberately, so it cannot cannibalise the tier above. And a useful detail if you are sizing it: failed runs, validation runs and design-time runs do not count against the automation quota — only scheduled and API-triggered executions do.
The wider point is that the company went private in 2024, which removed both the quarterly-earnings constraint that would have made a disruptive repackaging painful, and the public financial disclosure you would otherwise use to assess vendor viability. Both halves of that are worth saying in a procurement conversation.
What the analysts say, and how much to weight it
The principal quadrant report for this market published its 2026 edition in June. Microsoft was named a Leader for the nineteenth consecutive year and Qlik for the sixteenth; ThoughtSpot, Google and Salesforce are also placed as Leaders. The stated framing is a shift from dashboard-first BI toward AI-augmented, decision-centric analytics, with BI platforms positioned as governed front ends for enterprise AI.
On market share, be honest that there isn't a trustworthy number.
The figures that circulate — roughly 22% for Power BI against 18% for Tableau — come from web-technology detection. They measure installations detected, not revenue and not seats. The rank order is probably right; the percentages should not be quoted as share.
And market size estimates disagree by nearly two-fold — from around $27B to around $50B for overlapping years, across reputable firms. That is a definitional disagreement about what counts as BI, not a data disagreement.
Teach the spread rather than a number. Being the person who says ‘the estimates differ by 2× and here is why’ is more credible than the person with a confident figure.
Check yourself