Start with the only distinction that matters
Digital advertising gets taught as a list of platforms. That is the wrong starting point, because it makes six things look like six choices when really there are only two jobs, and every channel does one of them.
The first job is capturing demand that already exists. Someone needs a plumber tonight and types it into a search box. The second is creating demand that does not exist yet. Someone who will need a new roof in eight months has never thought about you and is not searching for anything.
Once you know which job you are trying to do, the channel question mostly answers itself. Get this backwards — asking display ads to close urgent business, or search ads to build a brand — and you will conclude that advertising does not work, when what did not work was the assignment.
Paid search
What it is: text ads on the results page when someone searches a term you bid on. You pay when they click, typically $3 to $60 depending on how much a customer is worth in your category.
What it is for: capturing demand. It is the highest-intent inventory in advertising, because the person told you exactly what they want, unprompted, seconds ago.
Its limit is that the ceiling is fixed. In a county of 60,000 people, there may only be 400 searches a month for what you sell. You can win all of them and still be nowhere near your growth target. That is not a targeting problem you can optimize your way out of — it is arithmetic, and it is the single most common reason a local advertiser plateaus.
Paid social
What it is: ads in the feeds of Meta, TikTok, and increasingly Reddit and LinkedIn. You pay per thousand impressions or per click, at costs well below search.
What it is for: mostly creating demand, sometimes capturing it. Nobody opens Instagram to buy a water heater, but a well-made ad about a $59 tune-up in front of homeowners in three ZIP codes creates a call that would not otherwise have happened.
The thing operators get wrong: judging social by the same standard as search. It will always look worse on last-click cost per lead, because it is doing the earlier, harder work. Judge it on whether total leads across the whole program went up during the flight.
Streaming and connected TV
What it is: your fifteen or thirty seconds inside streaming content, bought by audience and geography rather than by market and daypart. Non-skippable, full screen, on the living room television.
What it is for: creating demand and, just as importantly, buying credibility. A business that appears between episodes reads as established in a way no feed placement does. Businesses that add it commonly see searches for their own name rise, which is the clearest signal that it worked.
Cost: real household reach inside a county now runs a few hundred dollars a week. The old minimums that kept local advertisers out are gone.
Programmatic display and retargeting
What it is: banner and video placements bought automatically across millions of sites and apps. Cheap on a per-impression basis, and the most misused channel in local advertising.
What it is for: two things well and everything else badly. It is excellent at retargeting — staying in front of people who already visited your site and did not convert. And it is useful for broad, cheap presence in a defined geography when paired with something more persuasive.
Used as a standalone lead generator, it disappoints almost every time. If someone sold you display as your primary channel, that is why the results were thin.
Streaming audio and digital out-of-home
Audio reaches people during commutes, workouts, and chores — moments with no screen and no competition for attention. It is inexpensive, and for service businesses with a memorable offer it punches above its cost.
Digital out-of-home means programmatic screens: gas pumps, gyms, transit, retail, roadside boards. It is now bought by ZIP code and daypart in small budgets. What it does that nothing else does is make a small business look permanent.
Neither of these will be your first channel. Both are good third and fourth channels, and both are routinely skipped by providers because each adds work for the same fee.
What a sane first plan looks like
For a local business with a proven offer and $2,000 to $6,000 a month: put roughly 60 percent into demand capture, which for most means paid search plus local service ads. Put 30 percent into one demand-creation channel — paid social if your offer benefits from being seen, streaming TV if credibility is your obstacle. Hold 10 percent for testing something new every quarter.
Add retargeting the moment you have enough site traffic for it to matter, which is usually sooner than people think.
Then leave it alone long enough to learn something. Six weeks minimum per channel. Most advertising that "did not work" was turned off in week three, during the exact period every platform spends learning who to show it to.
The part no channel fixes
None of the above compensates for an unclear offer, a phone nobody answers, or a website that takes nine seconds to load on a phone. Advertising buys attention. What happens to that attention afterwards is a business problem, and it is where most wasted budget actually goes.
Fix the offer and the follow-up first. Then the channel conversation becomes straightforward, and mostly a question of budget.