The term gets used loosely enough that it has almost stopped meaning anything. Agencies apply it to whatever they already sell. Here is the actual definition, the two research steps that should happen before any budget is committed, and the specific numbers that tell you whether a campaign is working.

Four questions, answered directly and in the order a business owner actually hits them.

What Is Performance Marketing?

The defining feature is that every outcome is attributable to a source.

It is advertising where you pay for a measurable outcome rather than for exposure. A billboard costs the same whether nobody looks at it or thousands do. A search ad costs you when somebody clicks, and a lead form costs you when somebody submits. The spend attaches to an action.

That definition has a second half most explanations skip: the outcome has to be traceable back to the specific ad, keyword, or audience that produced it. Without that trace, you are buying actions you cannot attribute, which is just advertising with extra steps.

Channels that qualify include paid search, paid social, display and retargeting, affiliate arrangements, and email marketing where the list was built through measurable acquisition. Email marketing is the interesting case: the send costs almost nothing, so the discipline sits entirely in how the list was acquired. What unites every channel on that list is not the platform but the accountability.

Two things it is not. It is not a synonym for paid advertising — a brand awareness campaign bought on a cost-per-thousand-impressions basis is paid, but nobody can say what it produced. And it is not the opposite of brand building. Firms need both. They simply need to stop pretending the second one is measurable on a monthly timescale.

How to Perform Market Research

Research answers whether demand exists before you pay to chase it.

For a local service business, market research does not require a survey firm. It requires four questions answered with real data, and all four can be answered free in an afternoon.

Does anyone search for what you sell?

The Keyword Planner inside a Google Ads account shows search volumes for your services in your specific area, and you can access it without running a campaign. This is the single most useful free tool available, and skipping it is how firms end up advertising into markets with no demand. If the volume is genuinely low, no amount of budget fixes that — the research has told you something valuable before you spent anything.

What words do buyers actually use?

Your vocabulary and your customer’s rarely match. Practitioners say “orthodontic treatment”; parents type “braces for kids.” Good market research surfaces the second list, not the first. Google Trends helps here by comparing relative interest between two phrasings, which tells you which one to build around.

Who already holds the positions?

Search your main terms and record what appears. Which firms hold the paid slots, which hold the map listings, which hold the organic results. Competitors who have been bidding for years have already tested what works in your market, and their ad copy is a free education.

What is a customer worth to you?

This is the part done last and it should be done first. Without an average customer value, no cost figure means anything. A $200 acquisition cost is excellent for a practice with a $6,000 average case and ruinous for one selling a $150 service.

How to Perform a Market Analysis

Research gathers the facts. Market analysis turns them into a decision, and the difference matters because most firms do the first and skip the second.

A working version for a service business needs four things on one page.

1. Demand volume. Monthly searches for your services in your service area, from the Keyword Planner rather than a guess.
2. Competitive cost. What the top of the auction currently costs per click, which tells you the entry price.
3. Your economics. Average customer value, and what proportion of enquiries you convert.
4. Capacity. How many new customers you could actually serve next month without quality dropping.

Put those together with your conversion rate and the answer usually falls out. If the auction costs $15 a click, the landing page conversion rate is 5%, and you close a third of enquiries, a customer costs roughly $900 before fees. That is either obviously viable or obviously not, depending on point three.

The fourth point gets ignored most often and causes the worst outcomes. Generating demand you cannot serve damages reputation faster than no marketing at all. A market analysis that returns “yes, but not yet” is a useful answer, and any provider who never returns it is not really analysing anything.

The Numbers That Actually Matter

Every metric above the line is an input. Only the bottom two are outcomes.

Most reporting stops three rungs too early, well before return on ad spend enters the conversation. Here is the full ladder and what each level is honestly worth.

A word on attribution, because it is where most disputes start. It is the process of deciding which touchpoint gets credit when a customer encountered you several times before buying. Somebody who found you through a blog post, left, saw a retargeting ad, then searched your name and called has three claimants on one sale.

There is no perfect answer to this. Last-click is simple and undercredits everything upstream. Data-driven models distribute credit across touchpoints and require volume most local businesses do not have. For a firm spending a few thousand a month, last-click attribution is usually the honest choice — imperfect, consistent, and nobody pretends otherwise.

How to Measure Social Media Marketing Performance

Engagement is an input. Report it separately from outcomes.

This is the channel where measurement goes wrong most reliably, because the platforms surface engagement metrics prominently and conversion data reluctantly.

Separate the two kinds of number and the confusion clears. Likes, shares, comments, reach, and follower growth measure whether content resonates. They belong in a report, labelled as what they are: leading indicators. Enquiries, acquisition cost, and revenue measure whether the channel pays. Those are the outcomes.

Three practical points on social media performance specifically:

• Paid and organic need separating. Combined social media performance reporting hides which half is producing. A post that reached 40,000 people because $600 was behind it is a different fact from one that reached 40,000 organically.
• Credit assignment is weaker here than on search. Social discovery often precedes a branded search days later, which last-click will credit to search. Expect social to be under-credited and do not over-correct for it.
• Restricted categories cannot track lower-funnel events. Some industries lose access to conversion-based reporting on the major platforms entirely, which changes what honest social media performance measurement can even look like.

That last point catches firms out, and it is worth reading up on before committing budget — our breakdown of the platform rules that decide what is possible covers which categories are affected.

Putting It Together

The same four questions apply whether the channel is paid search or email marketing. The sequence that works: research demand before committing budget, analyse whether the economics close, launch with tracking installed before the first dollar rather than after, and report on the bottom two rungs of the ladder.

Most failures are sequencing failures rather than execution failures. Advertising launched before tracking. Budget committed before anyone checked whether search volume existed. Reporting built on clicks because nobody installed what was needed to measure anything better.

Two honest limits. Results vary by market, budget, and competition, and anyone promising a specific cost per acquisition before seeing your data is guessing. And measurement itself has real boundaries — attribution is an approximation, not a ledger, and a provider who presents it as exact is overselling.

If you want the search-side mechanics in more depth, our guide to what the conversion data says about landing pages covers the page that decides whether any of this converts.