Every engagement starts with read-only access and a few days of digging before we propose anything. This page is the actual checklist: what we look at, in what order, and why each thing earns its place.
Why read-only access first
The truth lives inside your ad accounts, analytics, and current systems, not in assumptions made from the outside. So we ask for read-only access up front, have a proper look, and then write the plan. You get our findings either way, and because the access is read-only there’s no risk to anything currently running.
Most audits take a few days. What comes back isn’t a scorecard; it’s a sequenced 30/60/90 day plan: what we’d fix first, what that unlocks second, and what we’d build third.
The checklist
Can the business see what happened yesterday?
This is always the first question: can you see, for any given day, every new lead or sale that came in and every ad dollar that went out, across all channels, in one place? If the answer is ‘sort of, across three dashboards’, then nothing downstream can be trusted. It’s why measurement is always the first fix. You can’t make good decisions if you can’t see today and yesterday clearly.
What number does the business actually run on?
Not the platform’s number. Yours. For an ecommerce brand, that’s usually revenue against total ad spend (MER), new versus returning customer revenue, and what a customer is worth over time. For a service brand, it’s the conversion rate from lead to customer and the lag between them. Once we know that today’s orders or leads become revenue at a known rate, spend stops being a leap of faith and becomes a forecast.
This is the most common gap we find, and it shows up on both sides of consumer. Service brands optimise on raw leads while the sales team knows most of them are unqualified, and nobody ever tells the platform which was which. Ecommerce brands optimise on add-to-carts or first orders while the real economics sit in average order value and repeat purchase. If the outcomes that matter aren’t being fed back through offline conversions or CAPI, the algorithm has spent months getting very good at finding the wrong people. Fixing this one loop often does more than any creative or budget change.
Frequency next to reach: is the account finding new people?
We pin frequency next to reach on every prospecting campaign, across every paid channel. Prospecting campaigns should sit under about 2.0 over a month. Above that, the campaign isn’t finding new people anymore; it’s showing the same ads to the same audience. Remarketing is the opposite, where frequency above 2.0 is the whole point because that’s where the ask gets repeated. When one campaign is doing both jobs, neither is being done well.
Are campaigns bidding against each other?
Overlapping locations, duplicated audiences, and no exclusions between prospecting and remarketing mean the account ends up competing with itself and paying a premium for it. We map out who each campaign is allowed to reach, then make those boundaries real with exclusions, keeping site visitors, engagers, and existing customers out of prospecting.
Delivery health: what is the algorithm telling us?
Learning Limited flags, long-running ads that have saturated their audience, and automation features switched on or off without a clear reason. We don’t panic-fix any of it. A long-running winner that still pays is left alone. But every flag goes on the map, because each one helps explain where results are leaking.
Does the message survive the click?
We cross-reference the message in every ad with the destination it lands on. If the hook promises one thing and the landing page opens with something generic, there is a disconnect. A single generic page catching every ad angle is one of the most common leaks we find, and one of the cheapest to fix.
Where does the conversion path leak?
For service brands, lead quality and lead volume sit on a seesaw: forms that ask too little send unqualified leads to your most expensive people, and forms that ask too much choke volume. We look at what the form asks, what the sales team says about the leads, and whether what sales learns ever makes it back into the system. For ecommerce, the same leak shows up differently: product pages that don’t answer the real objections, checkouts with friction nobody has measured, and offers that attract one-time discount hunters instead of repeat customers.
What’s the constraint that isn’t marketing?
Sometimes the biggest finding isn’t in the ad account at all: leads nobody has capacity to call, stock or fulfilment that can’t keep up with demand, a product that caps order value, or sales hiring that lags the pipeline. We’d rather tell you not to scale spend yet than make you grow too fast and break too many things. The plan gets sequenced around the real constraint, whatever it is.
The principle behind all of itFirst, do no harm. If something has worked for 18 months, it has earned some respect, so we build new structure around it and only replace it when the data shows the replacement is better. When an audit ends in ‘rip it all out and start again’, it usually serves the rebuild, not the result.
What you get
A written plan, usually within a week of getting access: the findings in plain language, the order we’d tackle them in, what each phase should unlock, and what it costs. If the biggest problem turns out to be something outside our work, like your sales capacity or your platform, the plan says that too.
From there, the engagement runs against that plan: an intensive first 90 days, then ongoing. 
