Media Buying: How to Compare Placements, Costs and New Channels

A media buyer turns an advertising plan into placements: choosing inventory, agreeing prices, launching ads and checking what the spend delivered. The work spans television, radio, print and outdoor advertising as well as search, social, websites, streaming services and AI conversations.

The hard part is comparing unlike offers. One seller quotes a price per thousand impressions, another charges for clicks, and a third bundles media with service fees. A useful buying decision makes those differences visible before the budget is committed.

Media buying vs. media planning

Media planning sets the audience, objective, channel mix, timing and budget. Media buying puts that plan into operation. Amazon Ads describes the distinction in those terms: the planner establishes the approach; the buyer secures and manages the placements.

A small team may do both jobs in one ad account. Changing a bid is a buying decision; changing the audience or the purpose of the campaign is a planning decision.

The main ways to buy media

A channel tells you where the ad appears. A buying method tells you how you obtain that placement. A billing model tells you what triggers a charge. These are separate choices: a native ad can be bought directly or programmatically, for example.

Programmatic does not always mean an open auction. Google’s programmatic guaranteed documentation describes deals with an agreed price and number of impressions over a defined period. Establish what is actually guaranteed rather than assuming that automation means either certainty or unrestricted access.

For the mechanics, read what programmatic advertising is. For a comparison of buying tools, see digital advertising platforms.

A media buying process you can use

1. Write the buying brief. Name the audience, offer, destination, objective, dates and maximum budget. For a lead campaign, define what makes a lead qualified before anyone reports a cost per lead.
2. Shortlist inventory. Ask where the ad can appear, how people encounter it, which exclusions are available and what reporting you will receive. Review a placement example rather than relying on a channel label.
3. Compare the complete offer. Separate media spend from platform, data, creative and service charges. Confirm minimum commitments and whether a quoted rate includes those costs.
4. Agree delivery and measurement. Record the buying terms, conversion event, reporting window and who can change or stop the campaign. For a direct buy, include what happens if delivery falls short.
5. Check the launch. Test links, creative, consent-aware measurement and the destination form or checkout. Confirm that campaign settings match the brief.
6. Review and reconcile. Compare delivered placements, qualified outcomes and invoices with the plan. Record what to repeat, change or stop before moving the next budget.

Amazon’s buying workflow includes insertion orders, creative trafficking, monitoring and reconciliation. Treat those as operating responsibilities even when a platform hides the paperwork behind campaign settings.

What does media buying cost?

There is no useful universal price for a media buy. Inventory, audience, geography, format, competition and the agreement all affect the offer. Start by making the unit explicit:

• CPM: media spend divided by impressions, multiplied by 1,000.
• CPC: media spend divided by clicks.
• Cost per qualified outcome: the costs you include in the comparison divided by outcomes that meet your agreed criteria.

Use the same cost basis across the comparison. A platform’s reported cost per lead may cover media alone; your acquisition decision may also need service and production costs. Neither figure is automatically wrong, but they answer different questions.

The ANA’s programmatic transparency research examines the gap between paid media cost and media quality. Its findings are a reason to inspect delivery and supply costs, rather than treating a low CPM as sufficient evidence of value.

A worked media buying example

Suppose a business compares two placements over the same reporting period. The following numbers are invented to demonstrate the calculation. They are not Thrad results, market rates or a recommended test budget.

Placement A looks cheaper by impressions and produces more leads. Placement B costs less per qualified lead: $2,400 divided by 16, compared with $2,400 divided by 8. If qualified leads are the objective, B deserves a closer look.

That does not establish a lasting winner. Check whether the leads become customers, whether the audiences overlap and whether the result holds as spend changes. The example shows why the denominator matters; it does not prove incrementality or predict the next campaign.

How to test a new media channel

Thrad’s perspective on early channel tests:

A new channel needs its own learning period. Comparing it immediately with a mature channel can miss part of its value. In conversational advertising, that value can include understanding what people say about a brand and the topics surrounding its buyers.

Give the test a learning brief alongside its performance target. The following is general editorial guidance, not a validated rule for scaling conversational advertising:

• Set two goals. Pair a commercial objective with a learning question. For example: can this placement bring qualified inquiries, and which product questions should our landing page answer?
• Put a limit on the learning. Agree a spend ceiling, review date and conditions for an immediate pause. A broken destination, incorrect tracking or unsuitable placement needs attention without waiting for the test to finish.
• Stay close to delivery. Review the available placement reports, creative and feedback with the channel team. Ask what can actually be observed; do not assume that every platform exposes the same data.
• Change something you can explain. If a message or destination appears mismatched, document the evidence and the revision. Avoid changing the audience, bid, creative and page together if you want to learn which change mattered.
• Decide what comes next. At the review date, ask what the test has taught you, what remains uncertain and whether another bounded test is justified. A decision to increase, change or stop spending needs evidence specific to your campaign; this framework supplies no proven scaling threshold for conversational advertising.

Keep observations and commercial outcomes in separate columns. Useful learning can justify a further test, but it does not turn a weak performance result into a strong one.

Check the reporting before judging the buy

Conversions can arrive after the initial visit. Google’s conversion-delay guidance recommends reviewing sufficiently mature date ranges so the report can include later conversions. Use your actual sales cycle and reporting lag when setting the review date; there is no single learning period that fits every advertiser.

Also check what counts as an attributed conversion. Google defines a view-through conversion as a conversion after an ad was seen without being clicked, and notes that a longer window can increase the number recorded. Keep that separate from click-based outcomes when comparing reports. Neither attribution method, on its own, proves that the ad caused the sale.

For a deeper evaluation plan, read how to measure AI-assistant ad performance.

Where conversational advertising fits in a media buy

Conversational advertising gives the buyer another placement context: a person working through a question in an AI interface. The brief needs to connect the offer to that task, with a landing page that continues it.

Thrad’s media-buying platform lets advertisers run campaigns across AI publishers with targeting, budgets and reporting managed in one place. Its DSP uses real-time bidding and conversational context. The self-serve platform lets advertisers set their budget, creative and targeting with no minimum spend.

For the agency-specific workflow, use our AI media buying guide.

Questions to ask before committing the budget

• Which placements or environments are included, and what can I exclude?
• What is guaranteed: price, volume, placement, timing, or none of these?
• Which companies participate in the supply route, and which fees are included?
• Who controls the account, creative, spend limits and pause decision?
• Which outcomes are reported, with what attribution rules and delay?
• What will we learn, and what evidence will support the next buying decision?

Supply-route visibility is a concrete operational issue. IAB Tech Lab’s Deals API is designed in part to clarify deal terms and the parties involved in curating and selling a package. Ask your buying partner for those answers in plain language before accepting a platform label as an explanation.