Premium Tutoring, Premium ROAS: How a UK Agency Brand Hit 3.11× Return on £1.07M in Ad Spend

This article is about a single brand within a larger engagement — the premium tutoring agency that consistently delivered the highest return on advertising spend of any brand in a three-brand EdTech portfolio. I'm writing about it separately because the strategy, the CRM logic, and the results were distinct from the marketplace and consultancy brands it sat alongside, and the lessons apply to anyone running paid acquisition for a premium, higher-ticket service.

The full portfolio story is covered in my separate article on the three-brand EdTech engagement. This piece focuses specifically on what made the agency brand different, why it outperformed, and what that teaches about running acquisition for premium services.

The Brand

A premium UK tutoring agency offering one-to-one private tutoring at university level, A-level, GCSE, and dissertation support. Higher ticket prices than the marketplace brand. More personalised service. A traditional agency model where the company matches students with carefully vetted tutors rather than allowing students to browse and self-select.

This distinction — agency model versus marketplace model — determined everything about the acquisition strategy. In a marketplace, the product is choice: dozens of tutors, transparent pricing, student-driven selection. In an agency, the product is curation: the company selects the right tutor for your specific needs, drawing on professional expertise to make the match.

Premium service, premium prices, premium expectations. The students and parents willing to pay agency rates were a different audience from marketplace browsers. They were less price-sensitive, more quality-conscious, more likely to research extensively before committing, and more likely to engage in a longer consideration process before their first lesson.

Why Premium Requires Different Acquisition Logic

The Marketplace Trap

The most common mistake in multi-brand portfolio management is running the same acquisition strategy across all brands. If the marketplace brand performance improves when you increase volume and reduce CPA, the instinct is to apply the same approach to every brand.

This doesn't work for premium brands. Here's why:

Volume optimisation kills quality signals. Google Ads algorithms, when given a volume objective, will find the cheapest conversions. For a marketplace, this is fine — you want maximum volume at minimum cost, and the marketplace's self-service model can absorb variable lead quality. For a premium agency, the cheapest conversions are often the least qualified. Price-sensitive leads who clicked because your ad appeared for a broad tutoring search are unlikely to convert at agency pricing.

CPA benchmarks are meaningless without ROAS. A marketplace might target a £20 CPA because the average customer value is £200. An agency might accept a £60 CPA because the average customer value is £800. Comparing CPAs across brands without adjusting for customer value leads to misallocation: the brand with the "worst" CPA might actually have the best return on investment.

Creative messaging serves different functions. Marketplace creative can emphasise speed, convenience, and access ("Find a tutor in 24 hours"). Agency creative must emphasise quality, expertise, and personalisation ("Your child's education deserves individual attention from a vetted expert"). Running marketplace messaging on agency campaigns attracts marketplace-quality leads at agency prices.

The Right Framework

For this agency brand, I designed the acquisition strategy around three principles that differ from the marketplace approach:

Optimise for ROAS, not CPA. Rather than targeting the lowest possible cost per lead, I targeted the highest possible return on ad spend. This meant accepting higher CPAs when the leads were more likely to convert at higher ticket values.

Quality gates before volume. Rather than generating maximum volume and qualifying downstream, I built quality signals into the acquisition process itself — specific landing pages for specific services, qualifying questions in the enquiry form, and premium messaging that self-selected for serious enquiries.

Distinct audience targeting. Rather than targeting everyone searching for "tutor" or "tutoring," I built audience segments around signals that predicted premium intent: specific university mentions, exam-specific search terms, subject-level specificity, and geographic targeting in affluent areas.

The Campaign Architecture

Google Ads Structure

The agency brand's Google Ads account was structured very differently from the marketplace brand, despite both operating in the same vertical.

Campaign segmentation by service level. Separate campaigns for university-level tutoring, A-level preparation, GCSE support, and dissertation assistance. Each service level had its own bid strategy, its own budget, and its own landing page. This granularity wasn't just organisational — it was strategic. University-level tutoring had the highest average customer value, which justified higher CPCs and more aggressive bidding. GCSE support had lower average values but higher volume, which required tighter CPA management.

Keyword quality over keyword quantity. The marketplace brand ran thousands of keywords across broad match, phrase match, and exact match variations. The agency brand ran a tighter keyword set focused on high-intent, service-specific terms. "Private GCSE maths tutor London" rather than just "maths tutor." "University dissertation help" rather than just "essay help." This tighter keyword strategy produced lower volume but higher average lead quality and conversion rates.

Negative keyword management. Extensive negative keyword lists filtered out traffic that indicated marketplace intent rather than agency intent. Terms like "cheap," "free," "online," "app," and "comparison" were negated early, preventing budget waste on leads who were looking for a marketplace experience rather than a premium agency service.

Ad scheduling. Attribution data revealed that the agency brand's highest-converting traffic occurred during specific hours — primarily evenings and weekends when parents researched tutoring options. Bid adjustments aligned budget with these high-conversion windows.

Landing Page Strategy

Each service level had its own dedicated landing page, designed around the agency value proposition rather than a generic tutoring offering.

Premium positioning. The landing pages emphasised the agency's curation process, tutor vetting standards, degree-level qualifications, DBS checks, and the personalised matching methodology. This wasn't marketing fluff — it was the genuine differentiator that justified the price premium, and the landing pages needed to communicate it clearly.

Trust signals. Trustpilot reviews featured prominently, but specifically reviews that referenced the quality of tutor matching and the personalised service — not just "great tutoring." DBS check badges, university partnership logos, and professional association memberships were tested in different positions and prominence levels.

Form design. The enquiry form was deliberately longer than the marketplace form. While the marketplace optimised for minimum friction (name, email, subject), the agency form asked for subject, level, specific learning goals, preferred schedule, and budget indication. This longer form served two purposes: it qualified leads before they entered the pipeline, and it provided the sales team with enough information to prepare a personalised response rather than a generic follow-up.

Mobile optimisation. Over 60% of the agency brand's traffic was mobile. The landing pages were designed mobile-first, with tap-friendly form fields, collapsible sections for detailed information, and click-to-call functionality for parents who preferred phone enquiry.

CRM and Lifecycle Automation

Agency-Specific Lifecycle Stages

The agency brand's CRM lifecycle was distinct from the marketplace pipeline, reflecting the longer consideration cycle and higher-touch sales process:

New Lead — enquiry submitted, qualification criteria pending
Qualified Lead — meets minimum criteria (correct geography, genuine interest, appropriate service match)
Consultation Scheduled — phone or video consultation booked with the agency team
Tutor Matching — consultation completed, tutor selection in progress
Trial Lesson — first lesson scheduled
Active Student — ongoing tutoring relationship
Retained Student — 3+ months of continuous tutoring

Each transition required specific conditions — no subjective judgments, no manual overrides. A lead didn't move from "Qualified" to "Consultation Scheduled" until a consultation was actually booked in the system. A student didn't move from "Trial" to "Active" until they'd completed their first paid lesson.

Nurture Sequences

The longer consideration cycle meant that nurture sequences were more important for the agency brand than for the marketplace. A parent researching premium tutoring might take 2-4 weeks from initial enquiry to first lesson, compared to 2-3 days for a marketplace booking.

I designed nurture sequences specifically for this timeline:

Immediate response. Automated acknowledgment within 60 seconds of enquiry submission, confirming receipt and setting expectations for next steps.

Value delivery. Within 24 hours, an automated email providing relevant resources — a guide to choosing the right tutor, what to expect from the first lesson, or tips for the specific exam they'd mentioned in their enquiry.

Social proof. At day 3-4, testimonials from parents in similar situations — another A-level maths parent, another university student seeking dissertation support.

Consultation invitation. At day 5-7, a direct invitation to schedule a free consultation, positioned as a no-commitment conversation about their specific needs.

Re-engagement. For leads who hadn't converted after 14 days, a re-engagement sequence addressing common objections: pricing concerns, uncertainty about the matching process, or questions about flexibility and scheduling.

The Results: Portfolio Context

The numbers for the agency brand, viewed in portfolio context:

Metric	Agency Brand	Marketplace Brand	Portfolio Total
Ad Spend	£1.07M	£1.63M+	£2.7M+ (student)
Attributed Revenue	£2.57M	£5.93M+	£8.5M+
Aggregate ROAS	2.41×	~3.64×	~3.15×
MQLs	26K+	56K+	82K+
Peak ROAS	3.11× (2024)	—	—

Why 3.11× Peak ROAS?

The 2024 peak ROAS of 3.11× for the agency brand was the highest single-year return of any brand in the portfolio. Several factors contributed:

Mature campaign data. By 2024, the campaigns had over two years of conversion data, enabling increasingly refined bidding strategies. Google's algorithms had enough signal data to distinguish high-value from low-value traffic patterns.

CRO compound effects. Two years of systematic A/B testing on landing pages, forms, and nurture sequences had compounded. Each individual improvement was modest (1-3% conversion rate lifts), but the cumulative effect of dozens of optimisations was significant.

Audience refinement. Two years of attribution data had revealed which audience segments, geographical areas, and search patterns predicted the highest customer lifetime value. Campaign targeting was progressively refined based on this revenue data, not just lead data.

Creative maturity. Ad copy and creative assets were evolved through systematic testing, with winning variations refined further rather than replaced wholesale. The creative portfolio became increasingly effective at communicating the agency's premium positioning.

Lessons for Premium Service Acquisition

Don't Compete on CPA

If you're running a premium service, your CPA will always be higher than budget competitors or marketplaces. That's not a problem — it's a feature. Your higher CPA reflects higher customer value, longer relationships, and better unit economics.

The metric that matters is ROAS, not CPA. A £60 CPA that produces a £300 customer is better economics than a £15 CPA that produces a £60 customer. But you can only make this comparison if you have end-funnel attribution connecting ad spend to revenue.

Build Quality Gates Into Acquisition

Don't generate maximum volume and try to qualify downstream. Build qualification into the acquisition process itself — longer forms, specific landing pages, premium messaging that self-selects for serious enquiries. You'll generate fewer leads, but each lead will be more likely to convert at your actual price point.

Separate Your Brand Strategies

If you operate multiple brands in the same vertical, resist the temptation to run identical campaigns with different brand names. Each brand has a different value proposition, a different audience, and a different conversion pathway. Each deserves its own strategy, its own creative, its own landing pages, and its own performance benchmarks.

Attribution Is Not Optional for Premium Brands

Premium brands can survive on gut feeling for longer than volume brands because each customer relationship is valuable enough to mask inefficiency. But attribution data reveals which campaigns produce your highest-value customers, not just which campaigns produce leads. Without attribution, you're optimising for volume when you should be optimising for value.

Related

Use Case: EdTech Tutoring Agency — Summary overview
Three Brands, Dual Pipelines — Full EdTech Engagement — The portfolio story
Revenue Systems — Attribution and reporting infrastructure
CRM Architecture — Lifecycle and automation

→ Book a Diagnostic Call — Let's talk about your premium acquisition strategy.