This guide is for a business owner who is considering Facebook and Instagram advertising, or already advertising without a clear system. It shows how the decisions connect into one chain, from the business goal to the result, and at every step it points to a deeper article. Updated: September 2026.
In short
- Facebook advertising does not suit everyone: if people are already actively searching for your product, Google search may be the more rational first channel.
- Three things have to be in place before you spend: an offer that persuades, a conversion path, and working tracking. Advertising does not fix what does not work without it.
- The campaign objective decides which people the system will look for. Traffic looks for clickers, not buyers.
- An audience is no longer a list of interests but boundaries and signals. The main levers now are the offer, the creative and your data.
- Platform figures are the start of the assessment, not proof. The result is what your own sales data confirms.
Contents
- Who this guide is for and what it does not solve
- When Facebook advertising is the right channel and when it is not
- Before the money: business goal, offer and conversion path
- The campaign objective: what the system will look for
- The audience: boundaries and signals
- Creative and message as part of the system
- Budget and cost: two parts and enough data
- Measurement: the Pixel, the Conversions API and consent
- What the platform figures prove and what they do not
- Testing without fooling yourself
- Retargeting: when it is justified
- The most common decision risks
- The next step: yourself, a contractor or an agency
Who this guide is for and what it does not solve
This guide is for the person who decides about money: the owner or the marketing lead. It is not for someone looking for which button to press in Ads Manager.
It does not solve three things. It does not set up your advertising, it does not tell you what your result will cost, and it does not promise that Facebook advertising will suit you. What it does instead is show the order in which the decisions have to be made, so that the answers to those questions come from your data rather than from guesswork.
The guide is arranged as a chain: channel, prerequisites, objective, audience, creative, budget, measurement, judging results, testing, retargeting, risks, next step. Each stage can be read on its own, and each one leads to the article where it is set out in full.
When Facebook advertising is the right channel and when it is not
Facebook and Instagram advertising reaches people who are not looking for you. That is its strength and its boundary at the same time.
If people are already actively searching for your product or service, an emergency plumber, a specific spare part, or accounting help with a clear need, Google search advertising meets them at the moment of searching. In ACCESSIO’s experience, in industries like these Google Search is often the more rational first channel, and Meta advertising comes in as an addition rather than a starting point.
Meta advertising can be the right first channel when demand still has to be created: the product is new, visually appealing, or the person does not yet know they have the problem you solve. Then advertising in the feed does what search cannot: it shows the product to someone who was not looking for it.
A more precise way to put it, and the one we use in conversations: Google search monetises intent that already exists, Meta creates new intent through the creative. For many companies the two channels complement each other. But if the budget allows only one, the choice of channel is the first and most important decision.
A simple self-test: search Google for your product or service the way a customer would. If competitors’ ads are already there and the search demand is steady, the intent exists in the market, and that is a signal to start with search. If the results are empty or unrelated, people are not searching yet, and that is Meta’s territory. This is a quick orientation, not a measurement: whether the demand is steady is shown by Google Keyword Planner or Google Trends data, not by a single search.
There are also situations in which Meta advertising is not the first step regardless of channel logic. They are collected in the table below, which is the central decision map of this guide.
| Situation | Why Meta advertising is not the first step | What to do first | When you can still start |
|---|---|---|---|
| The product is already actively searched for | People solve the need in a search engine, not in a feed | Google search advertising or local SEO | If the goal is to reach those who are not searching yet |
| The offer is untested | Advertising finds people but does not create desire for a product nobody wants | Sell manually, gather feedback from the first customers | If the product is new and a Meta video is itself the test |
| Tracking does not work | The system has nothing to learn from and you have nothing to judge by | Check the Pixel, the Conversions API and consent | Never, if the result happens on your website; if the result happens on the platform, website tracking is not a prerequisite |
| The economics do not work | If the profit from one purchase is small, the cost per result eats it | Raise average order value or repeat purchases | If the model deliberately rests on customer lifetime value |
| The problem is behind the advertising | A slow page, a weak form, slow replies to enquiries | Fix the page and the process | Once traffic from advertising finally reaches a deal |
More detail: How the choice of channel fits into the whole order of decisions is set out in our article on Facebook ads strategy.
Before the money: business goal, offer and conversion path
The third decision happens before Ads Manager is opened. In ACCESSIO’s practice the first conversation with a client does not start with campaigns but with the core business goals: what the company wants to achieve, and which principles everything else rests on, marketing included.
From the goal follows one specific result the advertising has to produce: an enquiry, a call or a purchase. Not “awareness”, but an event you can count and for which you can work out an allowable cost.
Then comes the offer. If someone who does not know you has no reason to act now, advertising will not fix that. A workable offer has a concrete benefit, an understandable first step, and a reason to act now rather than at some point.
And the conversion path: where the person has to arrive. In ACCESSIO’s practice the choice between a form inside the ad and the website is decided by how serious the purchase is. If the decision is simple, a form reduces friction; if the purchase is serious, the person has to reach a page where the content persuades them.
The third prerequisite is measurement readiness: whether the event you will judge the advertising by is being recorded in the account at all. Until it is measured, every conclusion about the advertising is a guess. More on that in the chapter on measurement.
The campaign objective: what the system will look for
The campaign objective in Ads Manager is not a formality. The system looks for people who are most likely to take the exact action you optimise for. If the objective is clicks, it will find clickers. If the objective is purchases, it will look for buyers.
In ACCESSIO’s practice one of the most common mistakes is a Traffic campaign where the business is waiting for enquiries or purchases. Clicks are cheaper, the numbers look livelier, and the results do not come. If your result is an enquiry or a purchase, a conversion-oriented objective such as Leads or Sales usually fits.
That does not mean Traffic has no place: it suits the stage where the action you actually want is reading. The mistake is expecting purchases from click optimisation.
The practical logic is simple: for a service business the result is usually an enquiry or a call, so Leads; for e-commerce a purchase, so Sales. If your first instinct is that the goal is awareness, ask yourself what will show you in three months that it paid off. If there is no answer, the goal is still enquiries or purchases.
The audience: boundaries and signals
A Facebook ads audience is no longer a list of interests you tick. The settings in Ads Manager now split into two groups with different meanings.
Boundaries (Controls) are not crossed by the system even with Advantage+ on: location, minimum age, languages and excluded custom audiences. Signals (Suggestions) are used as a starting point but expanded beyond when the system predicts a better result: included custom and lookalike audiences, age, gender, interests.
From that follows a practical rule: if something has to be strict, it has to sit in the boundaries. If you place something in the signals and expect only those people to see the ad, your expectation does not match how the system works.
At the scale of a small market this matters twice over: in September 2026 Ads Manager showed roughly 1.1 to 1.2 million people for Latvia from age 18. The pool is not large, so fine segmentation quickly becomes a problem, and in ACCESSIO’s practice it does more harm than good.
If you are unsure whether a split is justified, ask one question: will I show these two groups different ads? If not, there is no reason for the split. The most valuable thing you can have is your own data: a list of buyers with purchase values is a stronger signal than any combination of interests.
Creative and message as part of the system
Once the audience is signals rather than boundaries, the creative becomes the main targeting tool. The system watches which people respond to your message and shows it to similar ones. A message that speaks to everyone signals nothing.
So the message has to be framed as a testable hypothesis: who it is for, which obstacle it addresses, what it promises, and why it should be believed. One ad tests one hypothesis.
A practical note: in the mobile feed only the first part of the text is visible, roughly 125 characters, and the exact cut-off differs by placement and device. If you cannot state the core of the hypothesis in that length, it is not ready.
And when results do not come, the creative is the first thing worth checking: it is the easiest place for a mistake to slip in and the easiest thing to change. But change one thing at a time, or you will not know what worked.
More detail: The format of a creative hypothesis, with an example, is in the strategy article’s chapter on the creative hypothesis.
Budget and cost: two parts and enough data
Facebook advertising has no price list. You set the budget, and the cost per result is formed in an auction. So the question “what does it cost” has to be split in two: the ad budget you pay Meta, and the cost of the work, which always exists, whether it is covered by your time or by a provider.
The size of the budget is set not by market averages but by your economics: what one enquiry or purchase is worth to you, how many you want, and whether that budget, at a realistic cost per result, will produce enough events for the system to learn and for you to decide.
The technical minimum is not the answer: in August 2026 the interface required at least EUR 2.00 a day for a Traffic campaign and EUR 0.80 for an Engagement campaign in a euro account. That is the threshold at which advertising runs at all, not the amount at which it achieves its goal.
The minimum test plan is three sentences, written before you start: how many weeks we test, what number of events and what cost count as a signal, and what we will do in each outcome. That is enough for the data, not your mood, to make the decision.
Measurement: the Pixel, the Conversions API and consent
Meta sees only what your website sends it. Without tracking, a campaign can be optimised for clicks or other on-platform actions, but not for enquiries or purchases on your website, because the system has no event to optimise for and you have nothing to judge by.
The Meta Pixel is the browser signal, the Conversions API is the server signal, and Meta recommends using both. Together they need deduplication, or one event gets counted twice. Cookie consent is part of the tracking system: visitors who do not consent do not appear in the data, and that is not an error you can fix.
Whether tracking works can be checked without a developer: Events Manager, with six checkpoints. But a green status only means that data is arriving, not that it is reliable enough to decide on.
In ACCESSIO’s practice the most common situation is not an absence of tracking but tracking that was started and never finished: the Pixel counts page views, but the event that matters to the business, an enquiry or a purchase, is not defined or does not fire. The campaign then cannot optimise for what you need.
What the platform figures prove and what they do not
Ads Manager shows dozens of numbers, but not one of them on its own answers whether the advertising pays off. Each answers a narrow question: was the ad shown, did people click, what did one platform-recorded event cost.
The platform sees the first three layers: the impression, the click and the platform event. From the fourth layer onwards, whether the contact is real, whether the enquiry was qualified, whether a deal happened and what it brought in, only your business data answers. A platform Lead is not a qualified enquiry, and a platform Purchase is not an accounting sale.
The practical rule: before comparing two numbers, check that the period, the level, the result type and the attribution settings match. If even one differs, you are comparing two different measurements.
A good report is therefore not a table of metrics but a verdict: what happened in the period, what the data clearly shows and what it does not, how the platform figures relate to your business data, and what the next decision is. If a report leaves the interpretation to you, it has not done the hard part.
Testing without fooling yourself
Ads Manager will always show that one ad is cheaper than another. The question is what you are allowed to conclude from that.
Two ads running in parallel in one ad set are not a test: the system itself decides which to show more and to whom, and the audiences overlap. That is an observation, which can produce a hypothesis, but it does not prove that the creative caused the difference. A controlled test is one where the audience is split without overlap and only one element changes.
The second rule: a small volume of data does not justify a categorical conclusion. A difference of a few conversions can be noise. And if you change the budget, the creative and the audience at once, the result cannot be attributed to any of them.
Three names help you avoid fooling yourself: a controlled test, a parallel observational comparison, and a before-and-after observation. Only the first proves a cause. The other two give you a hypothesis, and that is plenty, as long as you do not call it proof.
Retargeting: when it is justified
Retargeting is advertising to people who have already interacted with your business. It sounds like the safest money in advertising, and that is exactly why it is overrated.
Retargeting is not one tactic but several, each with a different message: remove a specific obstacle for basket abandoners, give a concrete reason to act to those who viewed the offer, offer the next step to existing customers. If you show warm audiences the same ad as cold ones, retargeting loses its point.
It has two prerequisites that are often missing: working tracking and a large enough audience. In a small market the audience is often too small for a separate ad set, and in ACCESSIO’s practice a simpler structure, one broad ad set with existing buyers excluded, is usually better until the data has built up.
And retargeting figures almost always flatter it on the platform: people who were already close to buying would partly have bought anyway, but the platform credits the ad. The practical guide: pause it for a while and watch your total sales data. That is a before-and-after observation rather than proof, because the season and your other campaigns affect the result too, but for a small business it is usually the only signal available.
The most common decision risks
There are a few assumptions companies bring to Meta advertising that ACCESSIO’s experience has to correct almost every time.
The first: “managing ads is simple, we will do it ourselves.” The interface really is available to everyone. But the measurement architecture, the budget maths that lets the system learn, and the skill of telling a platform number from a business result take experience. Only the first click is simple.
The second: “interest audiences can be targeted as precisely as before.” They cannot. Interests are now a signal, not a filter, and the system expands beyond them when it sees a cheaper result.
The third: “a very cheap result is achievable.” Every stage of the funnel loses people: from seeing to clicking, from clicking to acting, from acting to a deal. A cheap click does not prove a cheap customer.
And when results disappoint, the order in which to look for the fault depends on whether the offer has already been tested. In ACCESSIO’s practice, where the offer is assessed at the start of the collaboration, we check the creative first, then the audience, then measurement, then the page, and the offer last. If you are doing everything yourself and nobody has assessed your offer, the order may be the reverse. With one exception: if you suspect tracking is broken, check that first, because otherwise every other number is unreliable.
The next step: yourself, a contractor or an agency
The decisions in this guide can be carried out in three ways, and each has its own cost.
Doing it yourself: no cash outlay, but a cost in time every week, and part of the budget goes on learning at the start. It is a sound choice if you have both the time and the willingness to maintain the skill.
An employee or a contractor: the cost is a salary or a fee, the gain is that the work continues in the weeks when you have no time.
An agency: a fixed monthly fee, a percentage of ad spend, or a combination. When comparing proposals, what matters more than the price is knowing what is included and what is charged separately.
When an external provider is the natural next step: when the prerequisites are met, the offer is tested and tracking works, but you have no time or wish to keep a hand on the campaigns day to day. If the prerequisites are missing, no provider replaces them, and an honest one will say so in the first conversation.
If you would like ACCESSIO to take on the decisions in this guide and the day-to-day management, the service page sets out exactly what management includes and how it works: Facebook and Instagram ads management.
Frequently asked questions
What does Facebook advertising cost?
There is no price list. You set the ad budget, and the cost per result is formed in an auction. The budget is worked out from your own economics: what one enquiry or purchase is worth and how many you need. The cost of the work is counted separately.
Does Facebook advertising work for B2B?
It can, if the offer can be put to someone who is not yet searching for it, and if tracking connects a platform enquiry to a real contact. A B2B sales cycle is longer, so a platform Lead is not yet a result; the result is what your sales data confirms.
How long before results appear?
The test period is measured in weeks, not days. The system needs enough events to learn, and you need enough to tell a trend from noise. If the budget is small, the same number of events simply takes longer.
Can I do it myself?
You can. The interface is available to everyone. The hard part is not the buttons but the measurement, the budget maths and the interpretation of the numbers. If you do it yourself, start with a tracking check and one clear hypothesis, not with ten ad sets.
Do I need the Meta Pixel?
Without tracking the system cannot optimise for enquiries or purchases, and you cannot judge the result. The Meta Pixel is the browser signal, the Conversions API the server signal, and Meta recommends both. You can check whether they work yourself in Events Manager.
Why does Ads Manager show more purchases than my shop?
The platform counts by the attribution window and the ad contact; the business counts by order date and fact. The platform also credits the ad with purchases that would have happened anyway, and it does not see cancelled orders. A gap is normal; an unknown gap is a problem.
In summary
Facebook advertising is a chain of decisions, not a list of settings. First the channel and the prerequisites, then the objective, the audience, the creative and the budget, then measurement and an honest assessment of the results. At every stage this guide points to the article where that step is set out in full.
If you would like an experienced provider to take on these decisions and the day-to-day management, see ACCESSIO’s Facebook and Instagram ads management service.
