150+ projects, 40+ clients, and the discipline behind them
Almost every one of those engagements involved moving a platform or wiring two systems together. 15 of them are named below with the work described — the ones we can talk about in detail. Then how this work is actually run, and who owns each part of it.
The track record, in one paragraph
The InHouse Marketers has delivered 150+ projects for 40+ global B2B clients, most involving a platform migration or a system integration. This page names 15 of them — 9 platform migrations and 6 system integrations. The migrations include Eloqua to Marketo, Pardot to Marketo, Marketo to Marketo, Marketo to HubSpot, HubSpot to Marketo, HubSpot to Pardot, Salesforce to HubSpot, and both HubSpot and Pardot onto Odoo — covering consolidation of several instances into one following acquisitions. The integrations span Segment.io, Twilio, Drift, Clay and n8n, plus consolidating two HubSpot instances into a single Salesforce org, and 5 of them are written up in full.
Every move, named
Clients appear by sector rather than by name. Where we have written the engagement up, the row links to it — the rest describe the work rather than dress it up.
| Move | Sector | What it involved |
|---|---|---|
| EloquatoMarketo | Aviation | Assets, operational programs, integrations, webinar programs and nurture, plus the website’s own data-collection layer rebuilt and version-controlled so the site never stopped capturing. Read the write-up → |
| PardottoMarketo | Web solutions | Several separate Pardot instances consolidated into one Marketo instance after a run of acquisitions — workspaces and person partitions per acquired organisation, and a sequencing decision about which instance moved first. Read the write-up → |
| MarketotoMarketo | Web solutions | Merging an acquired competitor’s instance into the parent’s, keeping each brand’s database, sending domains and assets separated inside a single licence. Read the write-up → |
| HubSpottoPardot | Security services | System analysis first, then a written plan of exactly what needed migrating. The operational programs were the work; HubSpot workflows split across automation rules, completion actions and Engagement Studio. Read the write-up → |
| HubSpottoOdoo | Authentication software | A platform with almost no marketing operations community behind it. Learning Odoo well enough to teach it came before any migration, and program logic ended up encoded in regular expressions. Read the write-up → |
| PardottoOdoo | — | The same shape of problem as the HubSpot move: a platform with no native equivalent of scoring, grading or Engagement Studio, so the logic has to be rebuilt in the tool’s own constructs rather than carried across. |
| MarketotoHubSpot | — | Moving from a granular platform to a more opinionated one. Marketo smart campaigns do not collapse into HubSpot workflows one-for-one, so some logic consolidates, some is deliberately dropped, and the decision about which is a conversation, not a mapping exercise. |
| HubSpottoMarketo | — | Usually a move made because scale or complexity outgrew the original platform. Workflows become smart campaigns, and the programme, channel and token structure Marketo expects has to be designed up front — HubSpot has no equivalent to inherit. |
| SalesforcetoHubSpot | — | Mostly a question of ownership rather than transfer: which system is master for which object and field, which direction each one syncs, and what happens to the reports built on the fields that move. |
| Move | Sector | What it involved |
|---|---|---|
| Segment.iotoMarketo | Aviation & robotics | Several data sources brought into one system — website behaviour through Segment into Marketo after normalisation, with reporting built out in Tableau and the site updated to carry the tracking. |
| Two HubSpot instancestoOne Salesforce | — | Two marketing databases writing into a single CRM without overwriting each other: field-level ownership, deduplication across both sources, and sync rules that make it obvious which instance a record came from. |
| TwiliotoMarketo | — | SMS as a first-class channel rather than a bolt-on — consent captured and stored per channel, delivery and reply events written back to the person record, and opt-out honoured across both systems. |
| DrifttoMarketo | — | Conversational capture wired into the database properly: chat-sourced people created with the right source attribution instead of arriving as “direct”, and routing that respects the same rules as form fills. |
| ClaytoHubSpot | — | Enrichment that writes back safely — only filling fields that are genuinely empty, holding a record of where each value came from, and never letting an enrichment run quietly overwrite something a human entered. |
| n8ntoPipedrive | — | Workflow automation into a CRM that was never designed as a marketing platform: idempotent writes so a retry cannot duplicate a deal, and error handling that surfaces a failed run instead of swallowing it. |
Six disciplines a migration has to satisfy
A migration can be technically clean and still lose the quarter, or delivered on time and quietly illegal. These are the lenses the work gets held against — in brief here, in full if you want the detail.
Discovery
Stakeholder recall gets you a hypothesis list. The instance itself will tell you what actually runs — last-activity dates, non-null field counts, the programs nobody can explain.
Read the detail →Sequencing
Most schedule risk is waiting in someone else's queue. DNS, DKIM and the sandbox refresh go in during week one, before scope is even signed off.
Read the detail →Release engineering
Marketing platforms are production systems and deserve the same treatment: version control, environment parity where you can get it, and a rollback somebody has actually rehearsed.
Read the detail →Data and consent
An opted-in flag is a boolean. Consent is a record — who agreed, to what wording, when — and it does not survive a migration by default.
Read the detail →Revenue continuity
Technical failure is loud and gets fixed the same day. The expensive case is the routing rule assigning into a queue nobody works, running unnoticed for six weeks.
Read the detail →QA and cutover
Clicking around finds the faults you already suspected. A matrix finds the others, and the parallel run is the only test that proves anything.
Read the detail →Before you commission one
How many marketing automation migrations have you delivered?
What actually breaks in a marketing automation migration?
How long does a platform migration take?
Can you migrate without campaigns going dark?
What happens to consent and suppression lists during a migration?
Do you handle integrations as well as platform migrations?
Bring us the messy version
The migration that stalled, the instance someone else half-built, the acquisition that left you with three databases. A free 30-minute call and a system audit will tell you what the work actually is — including the parts nobody quoted for.