Back to Blog

One structural fact shapes every local search decision. We measured 18 local service results pages across Canada and the United States and 17 of 18 led with a map pack, which occupies the first three positions before a single conventional listing appears. The same measurement found zero AI Overviews on those pages, so unlike categories saturated by generated answers, the click below the map is still genuinely available. The full data is published here.

Four questions that do the sorting

The fourth is the one most people forget to ask and the one that most affects the value of the spend. Work that builds assets you keep is a different purchase from work that rents visibility for as long as you pay. Both can be right. Buying one while believing you bought the other is not.

How the fee models differ, and what each one incentivises

Flat retainer. Predictable and easy to budget. The risk is scope drift in either direction: you can end up under-served in a quiet month or the agency can end up losing money in a busy one, which eventually shows up as reduced attention.

Percentage of ad spend. Common in paid media and structurally biased toward increasing spend. Not disqualifying, but you should know the incentive exists and ask directly how recommendations to raise budget are justified.

Project-based. Cleanest for defined deliverables like a site build or a brand system. Weakest for anything that needs sustained iteration, because the work stops when the project closes rather than when the outcome is reached.

Whichever the model, ad spend and management fees should be separated in writing. Ambiguity there is the single most common source of disputes.

Answers that should end the conversation

A ranking guarantee. Local results vary by the exact location of the searcher, so two people in the same city see different orderings. Nobody can guarantee a position in something that is not fixed.

Reporting that starts after the work does. Without a before, the after means nothing, and a baseline captured in week three quietly absorbs any early loss.

Impressions as the headline metric with no mention of enquiries. Impressions are an input and they can rise while nothing commercially useful happens. They are worth tracking, but a report that leads with them and never reaches calls, forms or bookings is managing your perception rather than your outcome.

No answer to "what are you not going to do." Every honest scope has edges. A provider covering everything for one fee is underscoping, overcharging, or has not thought about it.

Timelines worth setting straight at the start

The mismatch between these timelines is where most agency relationships sour. Organic work judged on a paid-media timeline always looks like failure at week six, and it is usually cancelled just before it would have started returning.

If you want to see how we scope this, our services sets out what each engagement covers, and the Calgary page explains how we work with businesses here. We also publish our research openly, including a 330-result study of who actually publishes the AI search tools category, so you can judge the thinking before talking to anyone.

Questions

What should a digital marketing retainer include?

A named scope of work, a stated cadence, named people doing it, and a reporting baseline captured before work begins. If a retainer cannot be described as a list of things that will be produced and when, it is selling availability rather than output.

Should ad spend be separate from fees?

Yes, and it should be stated plainly. Management fees and media spend are different things, and a percentage-of-spend model creates an incentive to increase spend. Neither model is wrong, but you should know which one you are in.

How do I know if an agency is doing the work themselves?

Ask who specifically performs each part and where they are. Outsourcing is not inherently a problem, but undisclosed outsourcing usually means the person you liked in the pitch is not the person doing the work.

What reporting should I expect?

A baseline captured before work starts, then consistent reporting against it. Leading with impressions while never mentioning enquiries is the most common way underperformance gets dressed up, because impressions can rise while nothing commercially useful happens.

Do AI Overviews change local marketing?

Not yet, in our measurement. Zero of 18 local service queries returned an AI Overview, while all 20 queries in a global software category did. Local search still returns a conventional results page led by a map pack, so the click below it remains available.

What is a reasonable timeline before judging results?

Paid channels produce signal within weeks. Organic and content work is a matter of months, because it depends on indexing, accumulation and corroboration. An agency that promises organic results in weeks is either misinformed or describing paid work.