Straight answers
The questions businesses actually ask about performance marketing, measurement and AI search, answered plainly and without the sales pitch. If your question isn't here, ask it directly: we answer those too.
Last reviewed: July 2026
Plans, buys and measures advertising that has to answer for revenue: paid search, paid social, creative testing, conversion work and the measurement underneath it all. That's the generic answer. Ours is stricter: measurement first, because optimising on numbers you can't trust is expensive guessing. We report in numbers a CFO believes, like MER and new-customer ROAS.
Only if the agency outperforms its own fee, which you can only prove with honest measurement. In-house wins on context; a good agency wins on reps across accounts, markets and platform changes. Plenty of clients run both. Audit first, then decide on evidence.
Straight answer: it depends on scope, and anyone quoting a price before understanding your business is guessing. What we can tell you is the shape. Every engagement starts with the Reality Check, a fixed-fee three-week audit, so your first cost is known up front. After that, a rolling quarterly partnership in three tiers matched to your stage. Fees agreed in advance, never hidden inside media spend.
Percentage-of-spend pricing rewards an agency for spending more, not for making you more. So we don't do it. Fixed price for the Reality Check, a set quarterly fee for partnerships, and your media budget goes to media.
The people you meet. Everyone here has more than a decade in the trade and current certifications across Google, Meta and Microsoft. No juniors learning on your budget, no account managers reading someone else's notes. That one's a refusal we publish.
Yes. Ad accounts, analytics, data, learnings: yours from day one. Holding accounts hostage is on our published list of refusals. Walk away whenever you like and everything comes with you. Anything else is a warning sign.
Inside a quarter you should see real signal, though not always where you expected it. The first 90 days run as a pilot with success criteria agreed and signed up front, so 'results' is defined before we start rather than argued about later. Measurement fixes pay back fastest; eliminating waste doesn't wait for creative to mature.
Yes, most of our work crosses borders. Cape Town base, clients in 42+ markets across Europe, the UK, the US and Africa, several languages between us. The platforms are global and evidence travels well. Our day overlaps European and US East Coast hours.
There's no hard minimum spend; there is a minimum seriousness. Engagements run from proving paid media can work at all to multi-market scale, and the tiers reflect that range. What actually matters: enough volume to make defensible decisions, and the will to act on what the data says. If paid media isn't where your next cent should go, we'll say so.
No, and be wary of anyone who says otherwise. What we sign instead: pilot success criteria agreed before a cent moves, and reporting you can check against your own revenue.
Three checks. Does reported performance reconcile with your bank account, or only with platform dashboards? Do you own your accounts and data? And can they explain, in plain words, what they changed last month and why? Reports full of clicks and platform ROAS while revenue stands still are their own answer. An outside audit settles it quickly.
Start with who does the daily work and how senior they are. Then ownership: whose name is on the ad accounts and the data? How do they measure incrementality beyond platform ROAS, and what happened the last time a test failed? Finally, ask for a result they're proud of, with the measurement behind it. Operators answer easily. Salesmen change the subject.
Our fixed-fee, three-week audit, and the start of every engagement. It goes through your tracking and measurement, your media accounts and your AI-search visibility, then puts the findings in writing: what percentage of your conversions are invisible, where budget is leaking and what to do about it.
A written report that's yours to keep: the truth about your tracking, where spend is being wasted, how visible you are in AI search, and a prioritised 90-day plan. Any competent team could act on it, including one that isn't us. No obligation to continue.
Because every optimisation inherits the quality of the data under it. Typically 15 to 30 percent of 'converted' customers would have bought anyway, and platforms mark their own homework. Fix measurement first and every decision after that, ours or yours, stands on evidence. It's also simply the cheapest performance gain available.
Three paths. Take the report to your own team; it's yours to keep. Fix the urgent items with us in a 90-day pilot, success criteria signed first. Or continue into a quarterly partnership sized to your growth. Most clients start with the pilot. Nobody is locked into anything.
Because platforms count generously. View-through conversions, modelled conversions and overlapping credit mean every channel claims a slice of the same sale; stack the claims and you've sold more than you invoiced. iOS widens the gap further. We reconcile against blended, bank-account numbers like MER and treat platform ROAS as a directional input, not truth.
MER is total revenue divided by total marketing spend. That's it, and that's the point: it can't be inflated by attribution tricks because it ignores attribution entirely. If MER holds while you scale, the business is actually growing. Platform ROAS answers 'did this ad account flatter itself'; MER answers 'did the business grow'. We report both. Decisions follow MER and new-customer economics.
Apple's App Tracking Transparency took away the user-level identifiers platforms used to connect ads to purchases. Now they model what they can no longer observe, and report the model as fact. The fix: first-party data, server-side tracking, conversion APIs, measurement that doesn't lean on one platform's guesswork.
It sends conversion events from your server to the ad platforms instead of relying on browser pixels, which ad blockers, browsers and iOS keep strangling. If paid media matters to you, yes, you need it: it restores the signal the algorithms learn from, with consent handled properly. Standard in our measurement work.
Some would have, yes. Typically 15 to 30 percent of 'converted' customers were coming anyway. That's the incrementality question, the most expensive one most advertisers never ask. Holdout tests, geo experiments and real baselines separate advertising that causes revenue from advertising that takes credit for it.
No. Last-click died years ago; most reporting never got the memo. It hands the credit to the final touch, usually branded search, and starves the channels that created the demand in the first place. Keep it as one lens if you must, but never optimise to it alone.
The usual suspects, roughly in order: measurement loss dressed up as performance loss (iOS, consent, ad blockers), creative fatigue, audience saturation as spend scales, pricier auctions, and existing customers re-sold as 'new' conversions until that well runs dry. Start the diagnosis at measurement, not in the campaign settings. Several of these look identical in a dashboard, and only one of them is actually about your ads.
Some published, source-backed examples: vehicle marketplace leads up 46% year on year. A debt-review firm's gross profit up 52% on a 16% spend increase. Supplement brand ROAS up 69% while spend fell 60%. A plus-size fashion retailer's ROAS up 59%, new customers up 17%, on essentially flat spend. Each one is documented, with its measurement source, in the case studies.
Client data belongs to clients, so studies are anonymised by industry. The figures come from client records rather than platform dashboards, and every study names its measurement source and window. Where a client has approved being named or quoted, we say so.
Be the easiest source to quote. That means direct answers near the top of pages, question-shaped headings, quantified claims, the same company details everywhere, structured data and an llms.txt file, with AI crawlers allowed in robots.txt. Then work on what others say about you, because assistants weigh third-party mentions and reviews as heavily as your own site. AI Overviews already appear on roughly half of Google searches, so the audience reading these answers is already there.
Generative engine optimisation: structuring content so AI systems like ChatGPT, Gemini and Perplexity cite or recommend you in their answers. It rewards clear, quotable, evidence-dense writing and machine-readable structure, and it's measured in citations rather than rankings. SEO's successor, for a world where answers replace links.
One kind, yes: the agency whose value was operating dashboards. The algorithms do the buying now. What's left is judgment, the part that was always hard: what to feed the machines, what to believe, where the money goes next. Hire for that.
Hand-tuned targeting is mostly gone; platform AI does it better when it's fed well. The work moved upstream: clean first-party conversion data, creative tested at the pace the machines reward, independent measurement that checks the machine's claims. Broad targeting with strong signals now regularly beats hand-picked audiences; see the plus-size fashion study.
Yes, with supervision. Analysis, creative iteration at test pace, anomaly spotting, research. AI is a power tool here, not a strategist: every recommendation that reaches a client is checked against your data by a senior human who signs their name to it. And yes, we practise what we preach: this FAQ is written to be read by machines as well as people.
Ask it straight, get it answered straight. Or skip ahead and ask the bigger one: is your marketing telling you the truth?
