Marketo Engage and Account Engagement Rebuilds

An instance with 612 smart campaigns and no naming standard is not a Marketo problem. It is a bookkeeping problem wearing a Marketo licence. The last one we rebuilt, a national account team inside a logistics group, had 148 campaigns still triggering against a form that had been removed from the website in 2019, and four separate definitions of what a qualified enquiry was, one per person who had ever owned the instance. We archived 60 percent of it in the first three weeks. Nothing broke, and the reporting became legible for the first time since 2020. Which platform you actually have a problem with: Marketo Engage and Account Engagement fail differently and the fix is not the same. Marketo instances rot through unbounded cloning: every campaign a fresh build, tokens unused, no channel taxonomy, so nobody can answer what a program cost. Account Engagement instances rot at the boundary with Salesforce: prospects sync in ways nobody documented, completion actions fire from four places, and the connector user quietly owns records the sales team thinks they own. We audit before we quote scope, and the audit is chargeable work, not a sales call. Program architecture: A channel and tag taxonomy tied to how finance already reports spend, so program cost and channel roll up without a spreadsheet reconciliation. Program templates with local assets and my.tokens, so a campaign gets cloned and retitled rather than rebuilt, which is the only durable way a small team runs many programs. In Account Engagement, business units drawn on the line that matters, usually domestic Poland against export markets, since that boundary decides consent handling and language, not org chart convenience. Operational programs come before campaigns: Data normalisation for country, region and job title. Lifecycle stamping with one owner per transition, modelled in the Revenue Cycle Modeler or in Account Engagement's stage fields, and an agreed rule for what happens when a person moves backwards. A subscription centre that records channel preference rather than a single unsubscribe flag. Hard bounce and role address handling. Consent stamping with source, timestamp and wording version, which under RODO and the Polish electronic communications act is the record you will be asked for, not the campaign report. Sync behaviour, spelled out: Field by field, which system wins, and what happens on conflict. Which Salesforce record types sync and which are deliberately excluded. What the connector user can and cannot overwrite. Form handlers where your own site's forms must stay under your developers' control, native forms where progressive profiling earns its place. Tracker domains on your own domain with valid certificates, because first party tracking is the only tracking that still works in Safari. Scoring lives here, but is designed separately: We implement scoring in these platforms as part of a rebuild, but we do not invent the model inside a build project. The model deserves its own evidence, its own backtest and its own argument with your sales director. If you want both, buy them as two engagements and let the first inform the second. What we do not do: Ongoing campaign production, email copy, landing page design, and attribution reporting where the opportunity data underneath it is known to be unreliable. We will say so plainly rather than build a Marketo Measure dashboard on top of a pipeline nobody closes properly. Not for: Teams looking to migrate from Account Engagement to Marketo as a solution to a process problem, which it is not. Companies running both platforms in parallel as a hedge, which we will not maintain. Instances where the marketing operations owner is a contractor leaving inside six months, because the taxonomy is only worth what its enforcement is worth after we go. Handover: The taxonomy document, the operational program specifications, the sync map, and a working session with whoever runs it next, in Polish or English.