One media stack across Google, Meta and TikTok, piloted by engineers.

8,000 MAD/month of media spend is the hard floor below which Meta and Google cannot optimise properly — fewer than 50 weekly conversion events leaves the learning phase open and the algorithm guessing. Including the 6,500 MAD/month retainer and an initial 1,500 MAD creative envelope, a realistic month-one entry ticket lands at 14,000 to 16,000 MAD all-in. Anyone selling you a media stack on 3,000 MAD is selling hope, not performance — the math does not work and the algorithm will not let it work.

It depends on the demand shape. If your category already has measurable search volume in Morocco — a brand people google by name or a generic term with non-trivial impressions — start with Google Search to capture high-intent demand first, then layer Meta for retargeting and discovery in month two. If your category is invented or emerging (a new SaaS, a new D2C brand, a new B2B service nobody yet googles), start with Meta Advantage+ or TikTok to manufacture awareness, then add Google branded Search once the brand term picks up volume. Most Moroccan retainers benefit from Google first, e-commerce launches benefit from Meta first.

Search first, always, for the first 30 to 60 days. Performance Max is a black box that needs a clean conversion signal to optimise — feeding it broken client-side data on day one mostly trains it to spend on placements you cannot inspect. The Webotic protocol is: Search on branded + commercial keywords, Enhanced Conversions for Web wired through GTM Server-Side, 30 days of clean signal accumulated in Google Ads, then layer Performance Max with asset groups segmented by category and tight budget caps. Brands that start on PMax on day one routinely burn 15 to 25 KMAD on irrelevant Display inventory before the signal stabilises.

Weekly performance signal is visible from day 7 to 14: CPMs settle, CTRs stabilise on the new creative variants, the bidding algorithm exits its learning phase on accounts above the 8,000 MAD floor. Real ROAS uplift — the +340% median figure — crystallises between day 60 and 90, once server-side tracking has fed the algorithm 8 to 12 weeks of clean signal and creative iteration has cycled 20 to 40 new variants. We do not contractually guarantee a ROAS number, but we publish a target band within 30 days of starting and we defend it weekly in the dashboard.

Because there is no honest way to pilot a media stack on broken signal. Client-side tracking alone, in Morocco in 2026, captures roughly 65% of real conversions — iOS ATT cut 15 to 20%, ad-blockers and Safari ITP cut another 10 to 15%. Smart Bidding, Advantage+ Shopping and Performance Max all optimise on the data they receive: feed them 65% of reality and they over-spend on the wrong audiences for the first 60 days. Most Moroccan agencies sell server-side tracking as a 'phase 2' upsell at month four because it is profitable on a per-deal basis. We refuse the phasing because the first three months of an engagement are precisely the months where signal quality matters most.

Both, but as separate engagements. The default Webotic retainer is direct operation: a named engineer runs the account end-to-end, holds the credentials, makes the bid and creative decisions weekly. For in-house teams that want to internalise the engineering, we run a separate enablement engagement — typically a 12-week program with two engineers shadowing yours, the GTM Server-Side container handed over with documentation, a weekly review call, and a clean exit when the in-house team is capable of holding the cadence. We do not do half-measures (the agency holds the password but the client makes the decisions) — that arrangement breaks accountability within six weeks every time.