Contents
In short. Start with the channel closest to existing demand that your team can actually handle. If buyers already look for a solution, test search, maps or the marketplace where they choose. If demand does not exist yet, you need a demand-creation channel, and its sales signal comes later. Direct sales and partnerships make sense when you can name your prospects and the deal pays for sales time. Before launch, fix one channel, a spending cap, a deadline and a stopping rule.
Decide how a buyer will first meet you, then pick the platform
An acquisition channel answers one question: how will a future customer first meet your offer? The platform, ad account or agency comes after that. Starting with “Google or Instagram?” skips the key issue: is the buyer already looking, or must they learn about the solution first?
Channels fall into three groups. The first meets existing demand: search, maps, marketplaces and aggregators. The second creates demand: social media, video, display ads and content. The third relies on direct contact: partnerships and referrals, direct sales and cold outreach, email and messaging to your own list. Each differs in signal speed, cost and risk.
Google search versus Meta ads is compared in “Where to find customers: search or social”; this guide comes one step earlier.
Buyers already search: start with search, maps or the marketplace
Search. Google Ads Help describes Search campaigns as text ads on search results that reach people while they search on Google. The key condition: buyers already put their need into words.
Google Keyword Planner estimates monthly searches before you spend and forecasts clicks and conversions for a given budget; forecasts refresh daily and use the last 7-10 days, adjusted for seasonality. Google notes that results depend on bids, budget, ad quality and location targeting.
Maps. When customers choose a provider nearby, a business listing can be the first channel. Google names three main factors for local results: relevance, distance and prominence, which reviews and links affect, and states there is no way to pay for a better local ranking. So a complete, accurate Google Business Profile comes first.
Marketplaces and aggregators. If buyers choose inside a platform, test there. Amazon describes Sponsored Products as cost-per-click ads for individual listings that may appear at the top of, alongside or within shopping results, with no monthly or upfront fees. Rules differ by platform and change, so check seller documentation at launch. Platform fees reduce customer contribution and with it the money available for acquisition.
No demand yet: you need channels that create it
If the product is new or buyers do not know the problem can be solved, searches will not come. Show the offer to people who fit your criteria and explain why it matters now: the job of social media, video, display and content.
The campaign objective tells the platform what to look for. Meta’s developer documentation lists the current objectives as awareness, traffic, engagement, leads, app promotion and sales. LinkedIn groups objectives into awareness, consideration and conversion, by stage of the customer journey. Google says Display and Video campaigns serve reach and brand awareness as well as conversions.
Content and SEO give the slowest signal. Google’s SEO Starter Guide says some changes take effect within hours, others in several months, and suggests waiting a few weeks before judging an effect. So content rarely works as the only first test when you need results this quarter.
You can name your prospects: partners, direct sales and messaging
Partnerships and referrals borrow trust that a partner or a happy customer already has, and work best with a shared audience. Before launch, agree on compensation, who owns the customer and how the source is recorded: a code, a dedicated form or a CRM field. Without source tracking you will not know in a month whether the channel works.
Direct sales and cold outreach make sense when you can list prospects by name and the average deal pays for sales hours. The main cost is people’s time, not media, and it belongs in acquisition cost.
Email and messaging to your own list give the fastest signal if recipients agreed to hear from you. Check the legal and technical rules first. In the United States, the FTC’s CAN-SPAM guide says the law makes no exception for business-to-business email and that opt-out requests must be honored within 10 business days.
Gmail requires SPF or DKIM from all senders and, for senders of more than 5,000 messages a day, SPF, DKIM, DMARC and one-click unsubscribe, with spam rates kept below 0.3%. Rules differ by country; this is not legal advice.
Seven questions to answer before you choose
Answer in writing. “We don’t know” is useful: your first step is to check that point, not to launch ads.
Scroll right to see all columns
| Criterion | Question to ask | What the answer implies |
|---|---|---|
| Existing demand | Do people search for it, and does Keyword Planner show it | Yes: demand-capture channels. No: demand creation or direct contact |
| Deal size and contribution | What remains from a customer after variable costs | Low contribution cannot fund manual selling; high contribution justifies partners and sellers |
| Sales cycle | Days from first contact to payment | Long cycle: judge qualified inquiries first, sales later |
| Product complexity | Can a buyer understand the offer from one ad | A complex product needs a conversation, demo or expert content |
| Sales readiness | Who answers an inquiry, how fast and what they record | If nobody can respond, any channel wastes budget |
| Test budget | What you will spend for an answer, not a profit | It must cover the signals the decision needs |
| Signal speed | First conclusion needed in weeks or in a quarter | A short deadline rules out content as the only test |
Which channel to test first in your situation
The matrix is an editorial method, not an industry rule. If you match two rows, start with the one that signals sooner and that your team is ready to handle.
Scroll right to see all columns
| Situation | First test | What to prepare | First signal |
|---|---|---|---|
| Buyers look for a provider nearby | Business Profile and local search ads | A complete profile, reviews, call tracking | Calls and inquiries from the profile |
| Demand exists, the product is clear | Search advertising | Keywords, a page that answers the need, inquiry tracking | Inquiries in the first days |
| Buyers choose inside a platform | Marketplace or aggregator ads | A strong listing, contribution math with fees | Orders and listing visits |
| New product, quick decision | Social or video with a clear offer | Visual proof, an easy way to reach you | Response and inquiries; sales later |
| Large deals, prospects can be named | Direct sales, partnerships, LinkedIn | Account list, first email or call script | Replies and meetings within weeks |
| A list that opted in to messages | Email or messaging to the list | A segment, an offer, an unsubscribe route | Replies within days of sending |
| Complex product, long evaluation | Expert content plus narrow search terms | Material answering buyer questions | Search sooner, content in weeks to months |
How much to spend: size the test from the decision
Size the budget from the decision you need to make, not from what you can afford to lose. Five steps before launch:
- Define a qualified inquiry: the buyer and request traits worth a conversation. Agree it with sales (see the guide on lead quality).
- Take the allowable cost per qualified inquiry from customer economics, as derived in the guides on CPL, CAC and ROAS and on allowable lead cost.
- Decide how many qualified inquiries the decision needs and multiply by the allowable cost: that is the cap.
- Set a checkpoint at half the cap and write down which result stops the test.
- Set the observation period from the sales cycle: inquiries can be judged quickly, sales only after the normal buying time.
Illustrative example: $5,000 for one answer
Illustrative example. A service business with a sales cycle of about 60 days has an allowable cost per qualified inquiry of $250, taken from its customer economics. For a first decision it wants 20 qualified inquiries, an editorial choice for illustration, not a statistical threshold. The cap is 20 × $250 = $5,000.
The checkpoint is half the cap, $2,500. The rule is written in advance: fewer than five qualified inquiries by then means each costs over $2,500 / 5 = $500, twice the allowable level. Delivery pauses and the team reviews the offer, targeting and follow-up instead of adding budget. By contrast, 10 inquiries at the checkpoint cost exactly $250 each, 3 inquiries about $833.
Sales conclusions wait until 60 days after the last inquiry, and even then the sample is small: 2 deals from 20 inquiries is 10%, 3 deals is 15%. One deal moves the result by five points, so show counts beside percentages. The guide on sales conversion explains how to compare groups by first-contact date.
To compare two variants inside a channel, a Google Ads experiment splits a campaign’s traffic and budget between control and treatment arms. Meta’s split-testing documentation asks for one variable per test and notes that larger reach, longer schedules or higher budgets tend to give more statistically significant results. A small first budget answers only one question: does the channel work at all?
What to do with the result: four decisions
Move to the next matrix row only once you know why the test failed, or the new channel repeats the old mistake: the same offer, the same slow follow-up, the same vague idea of a good customer.
Team assignment for the week: fill in the criteria table, pick a matrix row, calculate the cap and checkpoint, name who handles inquiries and set the decision date.
Limits: the matrix and procedure are an editorial method, the numbers are illustrative, and messaging rules vary by country and need legal review. This journal is published with the involvement of WEBNOSTRA, a marketing services company that runs such tests; the method works without hiring anyone.
Scroll right to see all columns
| What the test showed | Decision | Next step |
|---|---|---|
| The checkpoint rule fired, or the cap ran out without enough qualified inquiries | Stop | Find the cause before choosing the next channel |
| Inquiries arrive but do not qualify | Fix and rerun | Change the promise or the screening, not the channel |
| Cost per qualified inquiry is within the allowable level | Continue | Wait for sales over the sales cycle |
| Sales are confirmed | Scale | Raise the budget in steps and watch whether inquiry cost climbs |
Key decisions
- establish whether demand already exists before choosing a platform
- test one channel first, chosen by deal size, sales cycle and sales readiness
- write down the cap, checkpoint and stopping rule before the first dollar is spent
What to discuss with your team
- Demand is checked: search estimates and the platforms where buyers choose
- The seven questions are answered in writing; every “don’t know” became a check
- One first channel is chosen from a matrix row
- A qualified inquiry is defined and someone owns the response
- The cap, checkpoint, stopping rule and observation period are recorded
- For email and cold outreach, consent and sender rules are verified
Terms in plain language
- Acquisition channel
- How a future customer first meets your offer: search, maps, social, a partner, a sales call, an email
- Demand creation
- Reaching people not yet looking for a solution, before a search exists
- Qualified inquiry
- An inquiry that meets buyer and request criteria agreed with sales in advance
- Sales cycle
- Time from first contact to payment; it sets when a test can be judged on sales
- Stopping rule
- A condition recorded before launch that ends the test, such as qualified inquiries costing too much at the checkpoint
Sources
- Google Ads: Choose the right campaign typesupport.google.com · checked 31/10/2026
- Google Ads: Use Keyword Plannersupport.google.com · checked 31/10/2026
- Google Ads: About campaign experimentssupport.google.com · checked 31/10/2026
- Google Business Profile: Tips to improve your local rankingsupport.google.com · checked 31/10/2026
- Google Search Central: SEO Starter Guidedevelopers.google.com · checked 31/10/2026
- Yandex Direct: Unified Performance Campaign (in Russian)yandex.ru · checked 31/10/2026
- Yandex Direct: display campaign results on the ya.ru home page (in Russian)yandex.com · checked 31/10/2026
- Yandex Wordstat: help (in Russian)yandex.ru · checked 31/10/2026
- Yandex Business: help (in Russian)yandex.ru · checked 31/10/2026
- Meta Marketing API: Ad campaign reference, objective fielddevelopers.facebook.com · checked 31/10/2026
- Meta Marketing API: Split testingdevelopers.facebook.com · checked 31/10/2026
- LinkedIn: Objective-based marketingbusiness.linkedin.com · checked 31/10/2026
- Amazon Ads: Sponsored Productsadvertising.amazon.com · checked 31/10/2026
- Gmail: Email sender guidelinessupport.google.com · checked 31/10/2026
- ConsultantPlus: advertising without consent, Article 18 of the Russian Advertising Law (in Russian)consultant.ru · checked 31/10/2026
- FTC: CAN-SPAM Act compliance guide for businessftc.gov · checked 31/10/2026