Last updated: 11 July 2026
A DTC brand can post excellent Google Ads ROAS while total new-customer growth remains flat. Branded search, returning buyers and demand created by Meta, creators or retail distribution can all be captured by Google and credited as if Google created them. This shortlist favours agencies whose published models can support the harder conversation about incrementality, creative and customer economics.
Disclosure: Upscale Digital publishes this article and may be included in the shortlist. Treat its inclusion as a commercial conflict and verify its claims with the same scrutiny applied to every other agency.
Methodology and limits
The shortlist is intentionally narrower than the general UK guide. We looked for first-party evidence in ecommerce, retail, customer-value activation, creative or integrated search. We then assessed how each operating model could fit a DTC team; we did not claim that a public case study guarantees similar performance.
Upscale appears because its founder profile states ecommerce, margin-led bidding and Shopping expertise. Impression, Croud, Hallam and Found bring different combinations of creative, customer data, multi-market execution and integrated search. None should be selected without validating new-customer measurement on the buyer’s own data.
| Test | What to establish |
|---|---|
| Incrementality posture | Will the agency separate branded capture from demand creation and admit what platform attribution cannot prove? |
| Customer economics | Can first-order margin, repeat rate, refunds and payback period shape targets? |
| Creative throughput | Who produces hooks, edits, product imagery and landing-page variants? |
| Channel interaction | How are Google results interpreted beside Meta, affiliates, email and organic demand? |
| Scale constraints | Does the plan respect inventory, cash flow and cohort payback rather than maximising reported revenue? |
Evidence-based shortlist
Upscale Digital
Published evidence: Upscale’s first-party profile highlights ecommerce, Shopping, PMax and margin-led bidding. Its compact model can fit a DTC team with strong internal brand and creative resources that wants senior Google Ads attention.
Due-diligence focus: The procurement test is incrementality: ask for a reporting design that separates brand, returning customers and prospecting, plus a test for an account where PMax ROAS rises but blended acquisition does not.
Impression
Published evidence: Impression publishes a paid-media case study for supplement subscription brand Nourished that discusses first-party CRM data and cross-channel attribution. That is directly relevant to DTC teams dealing with repeat purchase and channel interaction.
Due-diligence focus: Treat the case study as evidence of an approach, not a forecast. Ask which measurement and creative resources are assigned and how cohort value would be validated before entering bidding.
Croud
Published evidence: Croud’s MPB case study describes connecting first-party customer segmentation to Search Ads 360 to prioritise higher-value customers. The example is relevant to mature DTC brands with usable CRM intelligence and enterprise activation needs.
Due-diligence focus: Ask what data volume, identity matching and engineering support the proposed approach requires. A sophisticated value strategy is not useful if the brand cannot refresh or quality-check the source data.
Hallam
Published evidence: Hallam’s paid-media model includes creative, analytics, paid search, Shopping and broader media planning. That breadth is relevant where the DTC bottleneck is creative or landing-page effectiveness rather than campaign administration.
Due-diligence focus: Ask for the planned creative production cadence, ownership of source files and the measurement approach for upper-funnel video. Cheap reach should not be used to decorate a weak acquisition story.
Found
Published evidence: Found combines PPC, PMax, feed optimisation, YouTube and CRO within a wider search proposition. It may suit a DTC brand where product-page discovery and conversion work need to sit beside paid media.
Due-diligence focus: Request one cross-channel decision example tailored to the brand. The team should show how it would avoid paying indefinitely for demand already secured by organic rankings or customer retention.
Build the scorecard around new-customer contribution
Start with contribution after product cost, discounts, fulfilment, payment fees and expected returns. Then decide whether repeat purchases are predictable enough to include. A lifetime-value number built from the best historical cohorts can encourage overbidding for new cohorts that behave differently.
Separate measurement layers. Platform reporting is useful for optimisation; blended marketing efficiency shows the whole business; cohort reporting tests customer quality; experiments estimate incrementality. A DTC agency should know which question each layer answers and avoid forcing them into one dashboard number.
Creative is an operating input, not a quarterly request. If YouTube, Demand Gen or asset-rich PMax is in scope, the contract should state who writes briefs, who edits, how many concepts and variants are expected, how usage rights work and what happens when performance fatigues.
Operating-model decision
| Situation | Likely model | Procurement emphasis |
|---|---|---|
| Brand-led demand capture | Search and Shopping specialist | Protect efficiency but report branded and returning demand separately |
| Prospecting-heavy growth plan | Creative plus media team | Regular new concepts, prospecting measurement and landing-page continuity |
| Subscription model | Customer-data specialist | Cohort value, churn, refund treatment and controlled value updates |
| International DTC expansion | Multi-market team | Localised feeds, creative, pricing and market-level contribution |
Questions for the delivery team
Use recent DTC cohort and channel evidence when asking:
- How will brand and returning-customer revenue be reported?
- Which costs define contribution for bidding decisions?
- How is repeat value estimated and refreshed?
- What would prove Google is capturing rather than creating demand?
- How many usable creative variants are included each month?
- How are Meta and Google results reconciled?
- Which experiment is feasible at our current volume?
- When would you lower spend despite a strong platform ROAS?
How to compare final proposals
For DTC, put new-customer reporting, creative throughput, cohort inputs and cross-channel interpretation into the scope. These are operating requirements, not optional strategy language.
A DTC baseline must separate brand and returning demand before scale claims are judged. Retain ownership of customer-value logic, accounts and creative files.
Worked procurement scenario
A subscription brand reports 600% Google ROAS but 70% of conversions are branded or returning customers, while blended new-customer acquisition is deteriorating. The first assignment is not to scale PMax. It is to rebuild the measurement view, define a conservative new-customer value and create a creative test plan. Upscale may fit if internal teams own production and data; Impression or Hallam may fit when creative and analytics need agency resources; Croud may fit when CRM activation and international scale justify enterprise complexity. The chosen team should state what cannot yet be attributed.
Ask the chosen agency to set an expiry date on every temporary acquisition assumption. Introductory discounts, launch cohorts and seasonal repeat behaviour can make new-customer value look stronger than it remains. Refresh the value and compare predicted payback with observed cohorts before increasing targets.
Sources
DTC-related agency evidence checked 11 July 2026 and used only to identify plausible fit, not forecast results.