Meta says you made X.Your bank account disagrees.

The number every ad platform shows you is ROAS, return on ad spend, and it is the most flattering number in the whole business. It is one figure per channel, and every channel claims the same sale, so Meta counts an order, Google counts the same order, and your email tool counts it a third time. Add them up and the platforms together claim more revenue than the store actually made. The number that reconciles is blended: total revenue divided by total ad spend across everything you run, sometimes called MER, marketing efficiency ratio. One number, tied to the bank, that no single dashboard can inflate.

Even a true ROAS hides whether you keep the money, because it is measured before costs. A 4x return on a product with thin margin after cost of goods, shipping, payment fees, and the discount code the customer used can still lose money on every order. The number that answers whether an order was worth winning is contribution per order: what is left after those costs are taken out. A store can have a ROAS it is proud of and a contribution per order that quietly bleeds, and only one of those two numbers shows up on the ad platform.

The next split most founders never see is new versus returning revenue. A store that looks healthy on blended ROAS can be buying every single sale twice, because none of the buyers come back and each month starts from zero. Splitting revenue into new customer and returning customer shows whether the marketing is building an asset that keeps paying or renting one that stops the day you stop spending. Alongside it, average order value and payback period matter: payback is how long it takes a customer to earn back what you paid to acquire them, and if that stretch is longer than your cash can wait, being profitable on lifetime value does not help you pay this month's supplier invoice.

The plain takeaway is a test you can run on any agency. Ask them to report blended ROAS, contribution per order, and new versus returning revenue. If all they can show is platform ROAS, they are showing you the store from inside the ad account, which is exactly the one vantage point where the money never reconciles with the bank.

There is no single right channel, there is a right order, and most local stores get poor results by switching on five channels at once with a budget that cannot properly feed even one of them. Spread thin, every channel underperforms, and the founder concludes that none of them work when the real problem was starting them all together. The useful question is not which channel, it is which channel first, and the answer follows from whether the demand already exists.

Google search and Shopping catch people who already want the product and are typing for it. That is the highest intent traffic there is, and for a store with real search demand it is usually the first paid dollar, because you are meeting a buyer who has already decided and just needs to find you. Meta and TikTok work the other way: they create demand rather than catch it, which makes them strong for discovery, new products, and visual categories like fashion, beauty, food, and lifestyle that do well in Singapore feeds. They also need creative that keeps working as audiences tire of it, and a store that can convert cold traffic that was not looking for you a minute ago.

Email and WhatsApp are owned channels, not rented ones, and that is the point people miss. Once someone has bought from you or signed up, reaching them again costs almost nothing, which is why owned channels quietly carry a large share of a healthy store's revenue rather than sitting as an afterthought behind the paid campaigns. SEO is the slow compounding channel: no cost per click, but months before it pays, and its job is to lower your blended cost per order over time rather than to deliver a sale tomorrow.

The local specifics change the plan in ways a copied US playbook ignores. Singapore is a small market, so audiences saturate faster and ad frequency climbs sooner, which means the same creative fatigues quicker than it would in a larger country. Delivery expectations are high and often assume free or cheap shipping, checkout needs PayNow and the common local payment methods or you lose the sale at the last step, and the same shopper will happily cross-check your price on Shopee and Lazada before buying. A channel plan that pretends the marketplaces are not sitting in the same buyer's other browser tab is incomplete before it starts.

The first order usually loses money or barely breaks even once you count what you paid to acquire the customer. The profit lives in the second and third order, where you reach the same buyer for little or nothing. A store that never earns a second order is on a treadmill: it has to keep buying every sale forever, and the moment acquisition gets more expensive, the whole thing tightens. Retention is not a nice-to-have layered on top of growth, it is the part of the model where the margin actually is.

Repeat rate is the health metric that tells you which of those two stores you are running. If only a small and shrinking share of customers come back, cold acquisition has to do all the work and gets more expensive as you scale into thinner audiences. Lift repeat rate even a little and you change what you can afford to pay for a new customer, because each one is now worth more over time, which in turn lets you outbid competitors who are only counting the first order.

Timing is the lever most stores leave on the floor. A replenishment product, coffee, supplements, skincare, pet food, has a natural reorder window, and the flow should land just before the customer runs out rather than on a random calendar schedule that arrives too early to matter or too late to save the reorder. A considered one-off purchase does not get reordered at all, so the move there is a cross-sell to an adjacent product rather than a nudge to rebuy the same thing. Getting the timing wrong wastes the one message the customer was actually open to.

Segment by behaviour instead of blasting everyone the same promo. Best customers, lapsing customers, and one-time buyers each need a different message, and sending them all the same discount trains your best customers to wait for a code they would have happily paid full price to skip. The real point is not send more email, it is that acquisition and retention are one budget. Money that keeps an existing customer is usually cheaper per dollar of revenue than money that finds a new one, and most stores under-fund it because it never shows up as a shiny new campaign to announce.

It is not either or, it is a question of timing. Ads buy traffic today and stop the moment you stop paying, so you control the tap and you feel the effect the same day. SEO costs effort now and pays months later, then keeps paying without a per-click cost, so the return arrives slowly and then compounds. Framing it as a rivalry between the two is the mistake; they answer different questions about when you need the revenue.

Ads are the right first move when you need revenue this quarter, when you are testing whether a product even sells, or when you are launching something with no search history for anyone to find. SEO is the right investment when there is real, repeatable search demand for what you sell, people typing the product category, best, and location queries, and when you can afford to wait for it. Its payoff is a lower blended cost per order over time, because a chunk of orders start arriving without an ad bill attached and take load off the paid campaigns.

For a store specifically, SEO is not blog posts, and treating it as a content calendar is why so many ecommerce SEO efforts quietly fail. It is the product, collection, and category pages doing their job: clear titles and descriptions, structured product data so the listing can show price and stock in the search result, and fast mobile pages that do not lose the shopper before they load. That is a different discipline from SEO for a services website, and an agency that only knows the blog-and-backlinks version will not move a store.

The honest sequence for most stores is to run both against one number. Use ads to prove products and generate cash now, put SEO in parallel so that a year from now a real share of orders arrive without an ad bill, and judge both on the same blended figure rather than letting each claim credit inside its own dashboard. The caveat worth saying plainly: if there is little search volume for your category, SEO will not save the store, and paid plus owned channels do the work. SEO is a lever, not a guarantee, and it only pays where the searches already exist.