Ask buyers who own a BI platform what they would tell their past self and the answers cluster away from the dashboard: check which cloud contract the tool arrives with, budget for a setup that is the real project, find out who builds the hundredth dashboard, separate problems in the BI tool from problems in the data underneath it, take the appearance of the output seriously, and do not read a long tenure as a recommendation.
The six lessons
The cloud agreement shapes the shortlist — check that yours points the right way
When buyers explain why one platform was the easier choice, the agreement they already had comes up repeatedly. One organisation describes itself as "a cloud company" that was "already paying a lot of money to Google", which made Looker the smoother path for licensing; another describes their platform as "mostly just included in the Microsoft suite" and therefore the "company standard default". The mismatch is where it bites. A buyer at a company that is "a Google house" found using a Microsoft tool "a little bit difficult and cumbersome" to pull from BigQuery. The implication — ours, not a buyer's — is that the licence comparison your procurement can already run may narrow this decision more than the feature comparison will.
Named in this context: Power BI · Looker · BigQuery · Snowflake
The setup can be a bigger project than the dashboard
Setup burden recurs across the lower-rated accounts in this corpus. One buyer rates a platform a 6 because "it looks a little complex to set up". Another describes a "complex setup" and a "tough time setting it up", concluding that because they "don't have that kind of Tableau guru person" they are "looking not to even use Tableau in the future". A third says their platform "looks a little complex to set up", which creates a usability barrier forcing a VP to defer to the technical team. A fourth, eight years in, says the tool "requires a lot of development on the back".
Named in this context: Tableau · Looker · Power BI
Find out who builds the second hundred dashboards
The single-operator pattern runs through the lower ratings. One buyer explains their platform is used "for dashboarding that our analyst on the team has set up dashboards for us" — and when asked whether it is easy to use, answers directly that it is not. Another defers to their technical team. A third names the absence of a "guru" as the reason they are not expanding. Against that, buyers who rate these platforms well describe the opposite: explorers that are "super easy to use, super intuitive" for non-technical business stakeholders, and dashboards a marketer can build without a technical admin. Our recommendation, not a practice buyers describe: ask during the trial who would build the next fifty dashboards, and what happens to them when that person is on leave.
Named in this context: Looker · Tableau · Power BI
The layer underneath rates higher than the layer you are shopping for
Measured across this corpus, the warehouse and data-platform layer — Snowflake, Databricks, BigQuery — averages a full point above the visualization tools that sit on it, and comfortably above what buyers give software generally. The BI layer sits close to that general average rather than above it. This is a comparison worth running rather than assuming: it was checked against the corpus baseline, not asserted from the BI numbers alone. Buyers describe the reason in passing. One credits their platform's "bread and butter" data warehousing rather than its charts; another praises "secure connections to all of our data"; several name the connection to a warehouse as the thing that works. The implication — ours, not a buyer's — is that a team evaluating the BI layer should separate problems originating in the visualization tool from problems originating in the data underneath it.
Named in this context: Snowflake · Databricks · BigQuery · Domo
How the output looks is a real complaint, not a vain one
Buyers raise the appearance of finished reports directly, including in lower-rated accounts. One says plainly that "the dashboards don't look good" and that everything is "not really smooth in terms of user experience". Another is "actively considering alternatives" because the "look and feel is dated" and resembles something from an earlier era. A third finds their platform falls short on "data visualization capabilities" compared with what a rival could do. On the other side, the buyers who praise visualization use superlatives about the output specifically — "best visualization control" — and a buyer arriving from one of those tools says "I missed the visual charts that we got with Tableau". When the audience for a dashboard is an executive rather than an analyst, this is the thing they see.
Named in this context: Power BI · Qlik · Tableau · Looker
A long tenure is not an endorsement
Several buyers describe running the same platform for the better part of a decade while rating it in the middle. One has used theirs for eight years because it is "trustworthy" and "well-known", and in the same breath says it "requires a lot of development on the back". Another describes eight years of use driven by the fact that "leadership wants dashboards". A third is actively transitioning away after years on the platform. These accounts show why tenure should not be read as satisfaction on its own, which is worth remembering when a vendor cites its own retention. The clearest single number on this page says the same thing: buyers shopping for a BI platform rate the one they already have well below what the category averages.
Named in this context: Tableau · Looker · Qlik
How buyers rate the platforms
Every figure is the average across verified buyer interviews, on published transcripts. The table is grouped by layer because that grouping is the finding — the data platforms underneath consistently outrate the visualization tools on top of them. Platforms too thinly rated for a number are named without one.
| Platform | Rating | What buyers describe |
|---|---|---|
| Snowflake | 8.3 | The warehouse layer, and the highest-rated well-sampled platform in this set. |
| Databricks | 8.1 | Lakehouse and engineering-led analytics, praised for handling scale. |
| BigQuery | 8.1 | Google's warehouse, and the reason several buyers land on the BI tool beside it. |
| ThoughtSpot | 7.5 | Natural-language querying with no SQL required; buyers raise pricing. Sample too thin for a precise figure, so rounded to the nearest half. |
| Power BI | 7.4 | Arrives with the Microsoft estate. Buyers name Excel familiarity, and miss designed output when they come from elsewhere. |
| Tableau | 7.4 | The visualization itself is what its buyers talk about. Cost is raised even by those who rate it at the top. |
| Looker | 7.2 | Self-serve once the model exists; setup and the Google tie are where its lower raters concentrate. |
| Sigma | 7.1 | Spreadsheet-shaped analysis on the warehouse; one buyer finds it easier to manipulate than the alternative. |
| Domo | 6.8 | Buyers credit the data warehousing under it and the ease of sharing cards across an organisation. |
| Qlik | 6 | The lowest-rated platform in this table. Buyers name dated look and feel, and several are considering alternatives. |
The stories behind the lessons
The counter-current: not every BI problem starts in the BI tool
Several complaints in these interviews sit at the boundary between the platform and its operating environment: data connections, technical ownership, setup skills, and who builds dashboards for everyone else. That does not make the platform irrelevant — buyers also complain directly about usability, output and capability. It does mean an evaluation should separate product limitations from the conditions around the product before assuming a replacement fixes both.
Want this read against your own stack?
Get my read →What this means for your BI evaluation
Across these interviews the BI decision extends well beyond the dashboard. The existing cloud agreement repeatedly enters the choice, while setup skills, dashboard ownership and the condition of the data underneath shape how the platform is experienced once deployed — that last being the layer buyers rate a full point higher than the tools sitting on it. Our recommendation, not a practice buyers describe: run the licence comparison before the feature comparison, and rebuild two existing dashboards during the trial rather than two new ones, so the test includes your real data, your output requirements and your operating constraints. For the head-to-heads behind these lessons, see Power BI vs. Tableau, Looker vs. Tableau and Looker vs. Power BI, and for the churn evidence Tableau alternatives.
Common questions
What do buyers wish they had known before buying a BI platform?
Six things recur across these interviews. Existing cloud agreements repeatedly influence which platform is easier to adopt. Setup can require more technical work than buyers expect. Lower-rated accounts repeatedly describe depending on a specialist to build dashboards. The warehouse and data-platform layer averages a full point above the visualization layer. Buyers directly penalise poor dashboard output. And several long-tenured accounts still give their platform middling ratings or describe considering alternatives.
Which BI platform do buyers rate highest?
Among the well-sampled platforms in this table the warehouse layer leads: Snowflake at 8.3, with Databricks and BigQuery both at 8.1. Among the well-sampled visualization tools, Power BI and Tableau lead at 7.4, followed by Looker at 7.2, Sigma at 7.1 and Domo at 6.8. Qlik is the lowest-rated platform in this table at 6. ThoughtSpot shows 7.5, but on a sample too thin for a precise figure — at that size a rating is rounded to the nearest half, so it should not be read as sitting above the two platforms below it. A higher figure here reflects how satisfied each platform's own users are, not a head-to-head test.
Why do buyers shopping for a BI tool rate their current one so low?
Buyers carrying an active BI search rate the platform they already run at 6.2, against 7.4 across the category and below what buyers give software generally. The same pattern appears in other categories on this channel. Their interviews also carry recurring complaints around setup, dashboards only one person can build, output that looks dated to the executives who see it, and fit with the rest of their data stack.
Should we just use the BI tool that comes with our cloud?
These interviews cannot answer that, but they show the existing cloud agreement repeatedly influencing the choice — and they show where the fit can become partial. Buyers whose data and licence sit in the same place describe it as the smoother path, in those words. Buyers on the other side of that line describe the same property as friction: one at a company that is “a Google house” found a Microsoft tool “a little bit difficult and cumbersome” for pulling from BigQuery. The question these interviews raise is not whether the bundled tool is good enough in general, but how much of your data and your audience sits outside the cloud it came from.
Get this research made for your stack
Choosing or replacing a BI platform? Do a 15-minute interview about your stack and get this personalized — what your data already sits in, who would build the dashboards, and which of these complaints would land on you.
Get my personalized read — 15-min interviewNo password · your interview is anonymized before it ever informs a page like this one.