Budgets, auctions and bid strategies

0:00 Budgets, auctions and bidding

Here's a pattern that wastes a lot of money. A team launches ten ad sets with small budgets. After three days, most show a learning limited warning, the costs bounce around, someone panics and edits everything, and the whole cycle starts again. The ads were never given a fair chance. In this lecture you'll learn how the ad auction really works, why good creative lowers your costs, the budget types and bid strategies, what the learning phase is and how to budget for it, and how to scale without undoing your progress. You'll also plan a seasonal push, the way you would for Ramadan, Eid or Black Friday.

0:47 The auction

First, the auction. Social platforms sell ad space through auctions, but the highest bid doesn't automatically win. Meta describes a total value that combines three things. Your bid: what you're willing to pay for the result. The estimated action rate: how likely this person is to take your optimisation action. And ad quality: based on feedback and quality signals. Other platforms use similar logic. So what does that mean in practice? A relevant, engaging ad can beat a higher bidder with a weaker ad, and pay less. Here's the key idea: good creative literally lowers your prices.

1:29 Budget types

Now budget types. A daily budget is your average spend per day. Platforms may spend somewhat more on some days and less on others, averaging out over the week. A lifetime budget is a total over a fixed period, useful for launches and time-bound promotions like Eid or Black Friday, and needed for ad scheduling on some platforms. Then there's where the budget lives. Campaign-level budget, called Advantage campaign budget on Meta, lets the platform move money between ad sets towards the best opportunities. Ad-set budgets give you strict control, useful for clean tests or when you must guarantee spend on a specific audience.

2:14 The learning phase

Next, the learning phase. After launch or a significant edit, the delivery system explores who responds, and performance is less stable. Meta, for example, indicates an ad set usually needs around fifty optimisation events within a week after the last significant edit to exit learning. Let's do the maths, the simple worked example. If your target cost per acquisition is twenty-five dollars and you need about fifty events a week, one ad set needs roughly twelve hundred and fifty dollars a week, or about a hundred and eighty dollars a day. So with twelve thousand dollars a month, you can properly fund only about two ad sets optimising for purchase.

3:02 If budget is tight

What if your numbers don't support that? You've got three options. Consolidate into fewer ad sets. Use a campaign budget so the platform concentrates spend. Or temporarily optimise for a higher-volume event, like add to cart, until purchases grow. And avoid frequent significant edits, like big budget jumps, targeting changes or adding new creative into the same ad set, because each one can reset learning. The calculator in the lesson text works this out for you: your target CPA, events needed and monthly budget in, and the number of ad sets you can support out. If it says one, believe it.

3:46 Bid strategies

Now bid strategies, using Meta's names; others are similar. Highest volume, called maximum delivery on TikTok, spends your budget to get as many results as possible. It's the default and the right start for new accounts. Cost per result goal, or cost cap, aims to keep average cost near your target, but may deliver less. ROAS goal aims for a minimum return, for e-commerce with reliable purchase values. And bid cap sets a maximum bid per auction, for advanced users. The rule of thumb: start with highest volume until you know what's achievable, then add controls if costs drift above break-even. Set a cost goal far below reality and delivery simply stops.

4:35 Split and scale

How should the budget split across the funnel? For a growing e-commerce brand, the majority usually goes to prospecting, new people, because retargeting pools only grow if you keep reaching new people. A smaller share goes to retargeting, sized to the audience, because a small audience can't absorb a big budget without extreme frequency. And ring-fence a test budget for new creative. Then scale carefully. Raise budgets in steps of roughly twenty to thirty percent and wait a few days. For a big jump, duplicate the winner rather than shocking the original. And watch whether CPA rises faster than spend, which tells you you're near your efficient ceiling.

5:22 Example: Pakistani electronics (illustrative)

Let's look at a realistic scenario with illustrative numbers. A Pakistani electronics store targets a cost per order of two thousand rupees. It runs eight ad sets, each spending three thousand rupees a day and getting one or two purchases daily. None exits learning. The team consolidates into two ad sets with twelve thousand rupees a day each, on a campaign budget, broad targeting and six creatives each. With more events per ad set, delivery stabilises and cost per order becomes predictable. Same total budget. Different structure. That's the power of giving the algorithm enough data to learn from.

6:05 Example 2: Jeddah perfume, Ramadan (illustrative)

Now seasonality. Costs often rise at peak moments, Ramadan and Eid in Muslim-majority markets, White Friday and Black Friday, Christmas in the UK and US, eleven eleven in some markets, because more advertisers compete. Here's how a Jeddah perfume brand plans a four-week push. Two weeks before Ramadan, it builds creative and warms up prospecting at normal budgets, so campaigns exit learning before costs rise. During Ramadan, it uses a lifetime budget with scheduling weighted towards evenings after iftar. In the last ten days and Eid, it shifts to gift bundles and delivery-deadline messages. And it keeps the ROAS goal loose early, tightening only if delivery is healthy.

6:52 Mistakes and measures

Common mistakes. Many small ad sets that never exit learning. Big, sudden budget changes that reset learning. Cost caps set below realistic levels, so almost nothing delivers. And ignoring seasonal price rises, then blaming the ads. How do you measure success? Each ad set gets enough optimisation events to exit learning, so check the delivery status column. Every budget change is logged with the date and size, and none happen during a test's minimum run. And cost per acquisition holds within your target range as the budget steps up. When it stops holding, you've found your current efficient ceiling, and that's valuable information too.

7:37 Recap and try this now

Let's recap. Auctions combine your bid, the estimated action rate and ad quality, so relevant creative lowers costs. Choose daily or lifetime budgets and campaign or ad-set budgets deliberately. Fund each ad set for roughly fifty optimisation events a week, avoid edits that reset learning, and start with highest-volume bidding before adding cost or ROAS goals. Scale in steps and plan seasonal pushes early. Here's your try this now. Put your own numbers into the calculator from the lesson text. How many ad sets can your budget truly support? If the answer is fewer than you run today, write down how you'll consolidate.