The ops layer under 200+ live workflows
I build CRM architecture, routing, attribution and reporting for B2B and fintech go-to-market teams on Salesforce, HubSpot, Zapier, GA4 and Google Tag Manager.
A manual GTM back office, replaced by 200+ workflows
I walked into a manual GTM back office, wrote the definitions down, built the CRM architecture on one cleaned dataset, and handed the predictable work to workflows. Monthly reporting that took 1–2 hours now finishes in under 5 minutes. I ran that stack at PingPong for four years.
Read the full case ›200+ workflows run on these systems. I stood up and owned each one:
CRM architecture
Routing & automation
Attribution & tracking
Reporting & governance
Clarity first,
automation on top
01 · Create operational clarity
I start every ops engagement by writing definitions down: what a lead is, when it becomes qualified, who owns it after handoff. Most fights I get called in to referee turn out to be two teams counting different things under the same name.
I keep one object model across Salesforce and HubSpot, so lifecycle stage lives in one place and qualification logic follows a written rule. Every routing branch has a named owner. Nothing gets automated until marketing and sales have signed the same definition sheet.
Definitions first, or every dashboard becomes an argument02 · Build reliable data foundations
Once the language is set, I make the data obey it: every field gets one owner, one source, and syncs one way. I enforce hygiene at write time.
I keep the canonical record in Salesforce, sync HubSpot one way, and land 6sense intent plus GA4 events on the same contact. Validation catches bad writes at the door, which is why one cleaned dataset can feed eight reports that agree.
Clean on write beats cleanup on Friday03 · Automate predictable processes
Only when a process is boring and well-defined do I let workflows take it over. Automating a fuzzy process just delivers the mess faster.
Salesforce Flow and HubSpot Workflows assign, enrich, score and follow up, with Zapier and Make bridging what the native tools can't reach. At PingPong I built that out to 200+ workflows on one shared architecture, and lead handoff dropped from about 12 hours to 1–2 minutes.
Automate the boring, escalate the ambiguous04 · Measure what matters
I measure last on purpose: a dashboard on top of dirty data is a confident way to be wrong. I only build reporting once the definitions and the dataset can carry it.
I wire GA4 and GTM to CRM outcomes, set UTM strategy before spend scales, and point Looker Studio at the cleaned dataset. The monthly report took 1–2 hours to compile, now under 5 minutes, eight reports with no reconciliation step, and an LLM pass flags signals worth a human look.
A report nobody argues with is the deliverableWhat I wire together:
the CRM, routing and reporting stack
sales stack
HubSpot
Pipedrive
Salesloft
& workflow
Zapier
Make
Dify
HubSpot Workflow
reporting
GA4
Looker Studio
Power BI
Tableau
enrichment
6sense
ZoomInfo
Need one number your whole team can trust?
Bring me your field map and your routing rules; that is where I start.