Migrated from Marketo to HubSpot - 6 month report on workflow gaps

Hi everyone. We completed a migration from Marketo to HubSpot for our marketing automation about six months ago. Overall, the experience has been positive for cost and usability, but we've hit some specific workflow capability gaps that required workarounds. I wanted to share a concrete breakdown for anyone considering a similar move.

The biggest adjustment was around program management and granular scoring. In Marketo, we had complex, nested program structures with scoring rules that could be finely tuned per asset or interaction. HubSpot's lead scoring is more linear. We had to rebuild our logic using a combination of:
- HubSpot native scoring (for basic email/webpage engagement).
- Custom-calculated properties (updated via workflows for specific, high-value actions).
- A separate microservice (AWS Lambda) that handles complex multi-event sequences and posts the calculated score back via HubSpot's API.

Here’s a simplified Terraform snippet for the Lambda part that we use to keep infrastructure-as-code:

```hcl
resource "aws_lambda_function" "hubspot_score_processor" {
function_name = "hubspot-score-processor"
handler = "index.handler"
runtime = "nodejs18.x"
role = aws_iam_role.lambda_exec.arn

environment {
variables = {
HUBSPOT_ACCESS_TOKEN = var.hubspot_token
SCORE_THRESHOLD = "50"
}
}
}
```

Other notable gaps we documented:
- **CRM Sync Reliability:** The sync with Salesforce is generally good, but we've seen delays (10-15 mins) for complex object updates during high volume, which didn't happen with Marketo's native sync. We now queue critical real-time updates via a direct API call.
- **Dynamic Content in Emails:** Marketo's "dynamic content" modules were more flexible. In HubSpot, we often achieve similar results using personalization tokens combined with smart rules, but it requires more template variants.
- **Operational Reporting:** Out-of-the-box reports are great for marketers. However, for engineering observability on workflow performance and cost attribution, we had to pipe webhook events to our own CloudWatch dashboards.

The trade-off has been worth it for us due to the consolidated platform and lower total cost, but it required some engineering investment to bridge these gaps. I'm curious if others have found different solutions or hit similar walls.

-- Amy