Looker Consultant in 2026: Scope, Rates, Deliverables

A Looker consultant is hired for one of three reasons: the numbers on the dashboards no longer match, the Explores have become slow and the BigQuery bill has followed, or the team wants to open the model to conversational agents and needs it to be right first. This guide covers what such an engagement should include, what it costs in the United States and the United Kingdom in 2026, what a competent consultant actually hands over, and the signals that should make you walk away.

What a Looker consultant does

Looker is not a charting tool; it is a semantic layer written in LookML on top of a warehouse, most often BigQuery. The work therefore sits in the model more than in the dashboards. A typical scope covers:

• LookML modelling: views, Explores, joins, measures, access grants, one definition per metric.
• Performance: persistent derived tables, aggregate awareness, caching policies, the queries that cost the most.
• The upstream layer: dbt models on BigQuery, tests, where business logic should live so that it is tested once and read by Looker.
• Dashboards and adoption: fewer, faster dashboards; training analysts to use Explore rather than exporting to spreadsheets.
• Agent readiness: descriptions, tightened Explores and verified queries so that Conversational Analytics or the Looker MCP server can answer correctly.

Engagements come in three shapes: a time-boxed audit, a build (a new semantic layer or a migration) and a monthly retainer for ongoing changes. Most teams need the first before deciding on the other two.

Rates in the US and UK in 2026

Published rates from established Looker shops give a clear picture.

Format	Published range	Source
Boutique hourly rate, US	$200 to $250 per hour	Shearwater Data, April 2026
Consultant hourly rate, US	$120 to $185 per hour	Lets-Viz, 2026
Consultant hourly rate, UK	£90 to £140 per hour	Lets-Viz, 2026
Monthly retainer	$1,500 to $3,500 per month	Lets-Viz, 2026

Hourly rates are only half of the equation. Put them next to what Looker itself costs. Third-party estimates place the Standard edition at about $66,600 per year and Enterprise at about $132,000 per year, with a Viewer licence around $400 per user per year. Mammoth's 2026 analysis puts initial LookML development at $20,000 to $100,000 and ongoing maintenance at $40,000 to $160,000 per year. None of these figures come from Google, which publishes no amounts, but they give the order of magnitude: the model costs as much to maintain as the platform costs to license.

What a good consultant delivers

Whatever the format, the deliverables should look like software engineering, not like a slide deck.

Pull requests on a branch, not edits in production. Every LookML or dbt change arrives as a reviewable PR with a description of what changed and why. You keep the ability to say no.

Tests. LookML data tests on the measures that matter, dbt tests on the models that feed them, and a way to run both before a merge. A consultant who changes a revenue measure without a test that asserts its value against a known total is guessing.

Documentation in the model. Labels and descriptions on the fields, a note per Explore on what it answers, a short README on how the project is organised. Documentation kept in a separate document goes stale quickly; documentation in LookML is read by analysts and by agents.

Measurement before and after. Query times on the Explores that were slow, bytes billed by BigQuery over a comparable period, the count of duplicated measures, the list of dashboards that disagreed. Without a baseline, "it is faster now" is an opinion.

A prioritised list of what was not done. An honest audit finds more than five days can fix. The report should say what remains, in what order, and what each item costs.

Red flags

Some signals are reliable enough to end the conversation.

• Hourly billing with no scope. "We will see how it goes" means the risk is yours. Ask for a deliverable, a duration and a price.
• No public repository. A LookML or dbt developer with nothing verifiable online is asking you to trust a CV. Look for real projects, with tests, in a public repo.
• No dbt, no transformation layer at all. If the whole answer is "we will do it in LookML derived tables", your business logic will end up untested and rebuilt at every query.
• Rebuild first. A consultant who proposes to redo the model before reading it has not understood that your users depend on the current one.
• No mention of BigQuery cost. In Looker, performance and billing are the same problem. Someone who does not ask for the billing export is not looking at half of it.
• Dashboards as the deliverable. New dashboards on a flawed model are new places for the numbers to disagree.

Hourly, retainer or fixed price

Hourly works for overflow engineering when you already have a lead who scopes and reviews the work. A retainer suits teams with a steady flow of small changes and no in-house LookML developer; the useful range is a few days a month, cancellable monthly. Fixed price fits anything with a definable outcome, starting with the audit: a defined duration, a defined report, a defined set of fixes on a branch, and the risk of overrun on the seller's side.

For a first engagement, fixed price is the safest for the buyer. It forces the consultant to scope before quoting, and it gives you something to compare against the published hourly ranges above.

FAQ

Should the consultant know dbt as well as LookML?

Yes, if your warehouse is BigQuery. Logic that is tested once in dbt and read by Looker is cheaper and more reliable than logic rebuilt in LookML derived tables. A consultant who works in only one of the two layers will push complexity into the other.

Can a consultant work in a different time zone?

For most of this work, yes. A European consultant covers the US East Coast afternoon and the UK working day live, and asynchronous PRs do not need meetings. What matters is a fixed weekly review slot with the person who owns the numbers.

Next step

Neuravoid sells the audit at a fixed price: $4,500 for five business days on your LookML project, your dbt project on BigQuery, or both, with a prioritised report and the fixes delivered on a branch. For ongoing work, the retainer is $3,000 per month for three days of a remote analytics engineer, cancellable any month.

See the Looker and dbt audit and the analytics engineering retainer, and read next: LookML developer for hire: fixed price beats hourly and Looker pricing in 2026.

Sources

• Looker consulting services, Shearwater Data, April 2026
• How much does a Looker consultant cost in 2026, Lets-Viz
• Looker pricing, Toucan Toco, June 2026
• Looker pricing, Mammoth, September 2026