Win the first second, then the budget.

Four jobs that only work together. It builds and manages the media — structure, audiences, bidding, budgets, Spark Ads, catalogue and Shop setup where relevant. It supplies creative at the volume TikTok consumes, because a feed this ad-led exhausts assets quickly. It fixes measurement so the algorithm is optimising toward revenue rather than form views, and so your reporting can be defended. And it runs a testing programme that isolates hook, proof and offer, which is what makes each month cheaper than the last. An agency doing only the first job will look busy and stall.

Enough that a consolidated structure can exit the learning phase and that there is a monthly batch of new concepts behind it — those two costs are the real floor, and they are related, because thin budgets spread across many ad groups produce data nobody can act on. Rather than quote a universal number, we model it from your average order value or lead value, your target cost per result and the conversion volume the account needs to optimise. If the answer is that another channel should come first, we will tell you that before you sign anything.

Existing brand film can work as a component, but on its own it usually underperforms here: the feed rewards footage that reads as a person rather than as an advertisement. The industry has moved accordingly — creator content now powers 44% of paid media creative and Emplifi measured UGC-driven conversions at 6.73x in Q1 2026. Our practical answer is a mix: creator-style assets from our UGC pipeline for volume and proof, plus polished cuts from video production where the brand needs scale.

More than most brands expect, and the number scales with spend rather than with company size. Hook rate decays as frequency climbs, so an account running meaningful budget needs a steady batch of new openings, not a quarterly shoot. We scope the count from your spend level, how many audiences and geographies you run, and how quickly your current assets are decaying — then commit to it as a schedule so planning is possible. Starting smaller is fine. Starting with no pipeline is the mistake that shows up in month three.

Because they are answering different questions. TikTok counts conversions it believes it influenced, inside its own attribution windows and with modelling applied; your store counts what was paid for. Add duplicated events, micro-conversions counted as purchases and browser restrictions and the gap widens. Our approach is to make the gap a known quantity: audit and deduplicate events, move critical conversions server-side, reconcile against the store on a schedule, and report TikTok, store and blended efficiency side by side rather than picking a favourite.

Spark Ads let paid delivery run behind an existing organic post — yours or a creator's — keeping the handle, the comments and the social proof attached. They usually outperform an equivalent asset uploaded cleanly, because the ad looks like the thing people came to watch, and the engagement stays on the original post. The requirement is authorisation from the account that published it, which we handle as part of creator contracting. We use them as a default wherever a creator or organic post exists, not as an experiment.

It depends on conversion volume rather than preference. With little conversion history, lowest-cost delivery gathers signal fastest and cost caps simply throttle learning. Once the account has consistent volume, cost caps give predictable unit economics and value-based bidding becomes possible where revenue values are genuinely clean. We migrate deliberately, one change at a time, with the learning phase protected and a written rollback rule — because the most common cause of a sudden cost spike is three strategy changes made in the same week.

Often, with two adjustments. First, optimise toward a signal that predicts revenue rather than raw form fills, which means feeding qualified-stage data back from your CRM so delivery learns what a good lead looks like. Second, expect the assets to carry more explanation: practitioner-led demonstrations and honest talking-head footage outperform polished brand statements. We are also willing to say when a category is a poor fit — if the audience is narrow and the cycle very long, your budget may work harder in paid search first.

As a merchandising problem as much as a media one. Feed quality decides what can be delivered, so we audit titles, images, availability and pricing, then build product sets that separate proven sellers from the long tail rather than letting one catalogue campaign average everything. Video Shopping Ads and creator content are planned together, since the same footage sells the product and populates the storefront. Reporting is on contribution after returns and discounts, because gross revenue on a returns-heavy category flatters a decision you are about to regret.

Weeks one and two: tracking audit and rebuild, account audit, first hook families briefed. Weeks two to four: a consolidated structure launched with enough budget per ad group to exit learning, Spark Ads live where posts exist. Weeks four to eight: structured ad testing, with hook, proof and offer isolated so we learn which variable matters. Weeks eight to twelve: paced scaling on proven ad groups plus horizontal expansion into new hooks and markets. Throughout: a weekly working session and a monthly written review ending in recommendations rather than a screenshot of the dashboard.

We agree the answer before it happens. Every budget increase carries a rollback band — the cost per result range within which we hold, and the point at which we step spend back to the last stable level while the creative queue catches up. Because fatigue is tracked as a metric (frequency, hook rate decay, cost drift), the response is usually a new batch rather than a panic. You also see the decision log, so every increase and rollback can be explained a month later with the numbers that drove it.

From your business objectives backwards. If the goal is purchases, we optimise for purchase events and let TikTok learn from clean revenue signal; if the goal is qualified leads, delivery is pointed at the CRM stage that predicts revenue rather than the click. Reach and video-view objectives have their place for launches, and we say so plainly instead of selling awareness as performance. Whichever objectives apply, the account is documented so any client can see why each campaign exists and what it is meant to drive.

Both, in a set ratio. Trends and sounds help TikTok users accept an ad as part of the feed, so a slice of every batch explores what is working culturally this month. The durable half is built on your own proof: demonstrations, customer stories, comparisons and offers that keep delivering after the trend passes. Chasing trends alone produces spikes and no compounding; ignoring them entirely makes the spots feel like display advertising bolted onto a social app. We report which half is winning so the mix can be tuned with data.

Closely, and it should be planned as one programme. Influencer partnerships create content and credibility; paid delivery turns the good pieces into reach you control, usually through Spark Ads on the creator's own post. Organic posting teaches you which hooks TikTok users actually respond to before you spend money finding out. Our influencer marketing and social media management specialists work alongside the buyers, so nothing is briefed twice.

Very little, and we help with the rest. Access to the TikTok Ads Manager account or a new one created in your name, access to your store or CRM so results can be reconciled, your product or service facts, and whatever footage and customer stories already exist. From there we run the audit, agree objectives and a testing plan, and set out what each stage costs. If our assessment is that other services or channels should come first, that is what the audit will say — and you will still own the analysis.

Businesses whose product can be shown. Beauty, food and drink, home, apparel, fitness, pet and gadget brands see the fastest results because a demonstration does the selling, and their customers are already on the app. Service businesses can work well too when a practitioner explains something worth knowing. The clients who struggle are the ones with nothing visual to show and a very narrow target audience — and we would rather tell you that during the audit than take a retainer to learn it together over a quarter.

It behaves differently. Search captures people already looking for solutions, so it converts quickly and stops growing when demand runs out. This channel creates demand: users who were not shopping see something that makes them want it, which drives incremental sales but takes better measurement to prove. In most marketing plans the two work together, with search harvesting the interest social creates. We are happy to explore that split with you in the audit, using your own numbers rather than a generic case study.

A weekly working session and a monthly written review. The review states what moved, what we changed, what we learned and what happens next — spend, cost per result, click and conversion behaviour, revenue or pipeline, plus how each batch of new concepts performed against the last. Every client sees the same three views: what the app reports, what your own systems recorded, and the blended efficiency of the business. Nothing in that report needs an agency to interpret it, which is deliberate.

We can, and shared measurement is the reason clients ask. Our paid social, paid search and analytics teams work from the same definitions, so blended efficiency is a number rather than an argument between reports. Buying this channel alone is equally fine: we will report honestly against whatever your other agencies are doing, share the methodology, and keep everything inside your own accounts so nothing is stranded if the roster changes.