Meta ads budget for small businesses: how much to start with

"How much should we spend on Facebook and Instagram ads?" has no single answer, but there is a method for setting a Meta ads budget for small businesses. You start from what a customer is worth, estimate what each lead will cost, and check the budget against how Meta's delivery system learns.

This post walks through that method with a made-up salon, explains daily and lifetime budgets, the learning phase, GST on Meta's invoices, and the numbers to read once your ads are running. Every cost here is an example: your own Ads Manager numbers are the ones to use.

Start from what a customer is worth

The most you can pay for a lead is the gross profit from one new customer multiplied by the share of leads who buy. If a customer brings ₹6,000 of gross profit and one lead in five books, a lead is worth up to ₹1,200 before the ads lose money. If you want to keep half of that as profit, aim for ₹600 a lead.

Next, estimate what a lead will actually cost. Meta's numbers link together:

• CPM is the average cost of 1,000 impressions.
• Link CTR is the share of impressions that led to a link click.
• Conversion rate is the share of clicks that became a lead or an order.

CPM ÷ 1,000 ÷ link CTR gives the cost of a link click, and dividing by the conversion rate gives the cost of a lead. At a ₹180 CPM and a 0.9% link CTR, a click costs ₹20; if 8% of clicks become leads, a lead costs ₹250. CPM and CTR change with the audience, the ad, the placement and the season, so take them from your own Ads Manager history, or run a small test for a week or two first.

Our Meta Ads Budget Calculator does these sums. Enter what a customer or order is worth, your margin, your close rate, and your CPM, link CTR and conversion rate, and it shows the cost per lead you can afford, the daily budget for the leads you want, the impressions, reach and clicks that buys, and the profit to expect. It also checks your plan against the learning phase, and has a box to add GST if you cannot claim it back.

Daily or lifetime budget

Meta gives you two ways to set a budget:

• A daily budget is the average you are willing to spend each day. Meta can spend more than the daily budget on some days and less on others, but its help pages say that over a calendar week, Sunday to Saturday, it will not spend more than seven times the daily budget.
• A lifetime budget is one amount for the whole run of the campaign or ad set, with an end date, and Meta spreads the spending across those dates.

Our suggestion: a daily budget for lead ads you run all year, and a lifetime budget for a sale with fixed dates, such as a festival offer.

Keep the budget in as few places as you can. Meta's page on reducing cost per result suggests combining similar ad sets and letting a campaign budget share spending across ad sets, and warns that a sharp budget increase can push costs up.

The learning phase: why about 50 results a week matters

When you create an ad set, or make a significant edit to one, it enters a learning phase while the delivery system works out how best to show it. Results are less stable during this time.

Meta's page on Learning limited says an ad set is shown as "Learning limited" when it is unlikely to get around 50 optimisation events in the week after its last significant edit. Meta says this is not a penalty. Its suggested fixes are to combine ad sets, broaden the audience, raise the budget or the bid, or optimise for an event that happens more often.

What counts as a significant edit? Meta's page on the last significant edit lists pausing the ad set and changing the optimisation event, the audience or the creative; budget and bid changes may count too, depending on their size. So frequent big edits keep an ad set in learning.

A rough way to see what 50 results a week costs is 50 × your cost per result. This is our estimate, not a Meta formula. At ₹250 a lead, 50 leads a week is ₹12,500 a week, or about ₹1,786 a day. Many small businesses start below that. That is fine for a test: accept "Learning limited", keep to one ad set, and leave it alone for at least a week before you judge it.

Choose the objective that matches the result you want

Ads Manager asks for an objective first, such as Awareness, Traffic, Engagement, Leads or Sales. The objective decides what Meta counts as a result, and so what it optimises for and what your cost per result means.

• Leads suits a service business that wants enquiries, for example through an instant form that opens inside Facebook or Instagram.
• Ads that open a WhatsApp chat work with the engagement, leads, sales and traffic objectives, according to Meta's page on ads that click to WhatsApp. Our post on WhatsApp Business for shops covers setting up the chat side.
• Sales suits an online shop, and needs tracking on your site; for example, Meta counts landing page views only with the Meta Pixel installed.

GST on Meta's invoices

Meta's help page on GST and TDS for ads in India says ads bought in India are sold by Facebook India, billed in rupees, and subject to GST and TDS. GST is added to your ad charges, and you can add your GSTIN in your ad account's payment settings when the business country is India.

The Meta pages we found do not state the GST rate. Our calculator assumes 18% and asks you to check the rate on your Meta invoice. If you are registered for GST and can claim the input tax credit, the GST may not be a cost to you; your CA can confirm. This is general information, not tax advice.

Reading your results

Once the ads run, a few numbers tell you most of the story:

• Cost per result is the amount spent divided by results. Meta notes that some results may be estimated.
• Frequency is impressions divided by reach, an estimate of how often each person saw your ad. Meta's frequency page says that if results start to drop as frequency rises, the audience may be getting tired of the ad.
• Ad relevance rankings (quality, engagement rate and conversion rate) compare your ad with ads competing for the same audience. Meta's page on ad relevance diagnostics says they show only for ads with at least 500 impressions and are not used in the auction.

Ads Manager counts results with its own attribution settings, so also count enquiries and bookings in your own records. The number that matters is cost per customer.

When you have a few weeks of data, export a report from Ads Manager and open it in our Meta Ads Results Analyzer. It finds spend with no results, compares each row's cost per result with the median of your own rows, spots frequency rising while link CTR falls, lists Learning limited ad sets and rankings below average, and gives a pause, fix or scale list ranked by the rupees at stake. Its thresholds are our measures, and it reads only the file you give it.

Worked example: Glow Studio Salon, Salt Lake, Kolkata

Glow Studio is a made-up salon that wants bridal and party make-up bookings. All numbers are examples.

1. Value: an average booking of ₹12,000 at a 50% margin is ₹6,000 of gross profit. One lead in five books, so the break-even cost per lead is ₹1,200, and the owner aims for ₹600.
2. Expected cost: a two-week test gave a ₹180 CPM, a 0.9% link CTR and 8% of clicks becoming leads: ₹250 a lead, well inside the target.
3. Budget: 40 leads a month at ₹250 is ₹10,000, about ₹329 a day.
4. Profit check: 40 leads should bring about 8 bookings, so each booking costs about ₹1,250 in ads against ₹6,000 of gross profit. If the salon cannot claim GST, ₹10,000 becomes ₹11,800 at 18%, about ₹1,475 a booking.
5. Learning phase: 40 leads a month is about 9 a week, far short of 50, so the ad set will show Learning limited. The owner runs one ad set, makes no edits in the first week, and judges it on bookings after four weeks.

Before publishing, the owner pastes the ad text into our Meta Ad Policy Checker. It marks the first draft, "Want to hide your dark spots on the wedding day?", as needing care, because Meta's personal attributes policy does not allow ads that assert or imply things about the viewer. "Bridal make-up trials this month in Salt Lake" says the same thing safely. Meta still reviews every ad and makes the final decision.

Mistakes to avoid

• Splitting a small budget across many ad sets. Each one learns more slowly; Meta suggests combining them.
• Editing every day. Significant edits restart learning.
• Judging after two days. Give an ad set at least a week (our suggestion).
• Counting leads instead of customers. A cheap lead that never books is not cheap.
• Forgetting GST. Add it to your cost if you cannot claim it back.
• Doubling the budget overnight. Meta warns that a sharp increase can raise costs; raise it in steps.

Next step

Put your own numbers into the calculator, start with one ad set at a budget you can hold for a month, and review it weekly. If you also run Google Ads, our post on a Google Ads budget for a local business uses the same work-back method. Ads work better when people who check your profile find recent posts, so see our 30-day Instagram plan too. The tools here are free on our social media tools page, and none of them connect to your ad account. If you want help setting up your first campaign, talk to us.