George Maramigin

Attio Expert /

[George Maramigin @emaramigin]

I do the Attio Implementation with the AI.

A customer asked me if Attio connects to QuickBooks. Yes, it does, but via middleware like Zapier.

For the last 2 weeks, I've been playing around and building it:
https://t.co/Z1sNyElSav
Soon on Attio's Marketplace

#buildinpublic

[George Maramigin @emaramigin]

@sundv0 @Kellen_the_man @attio Before you move, map Pipedrive’s custom fields, activities, and owners to Attio objects before the import. I’ve handled these migrations, and reporting gets messy when the data model is decided afterward.

[George Maramigin @emaramigin]

@polsia Curious how it handles cross-object dedup, like a company record and a person record both pointing to the same domain but entered years apart. That mismatch causes more reporting damage than plain duplicate contacts ever do.

[George Maramigin @emaramigin]

@thebrettjansen Fastest way to run that audit: pull fill rates per field. Anything under 20% populated after 6 months isn't a data problem, it's proof nobody's process ever needed it.

[George Maramigin @emaramigin]

@Wilbur_Stein11 @attio Attio stays simple when the data model is clean and each automation has a clear owner. Otherwise, edge cases turn into manual cleanup.

[George Maramigin @emaramigin]

@PipelinePlayBK Most of that lost context isn't notes, it's the reasons a deal stalled or an objection got raised in a call that never got logged anywhere structured. Task closes, reasoning dies with the rep who had it in their head.

[George Maramigin @emaramigin]

@ctwtn Migration pages convert even better than alternatives pages. Someone searching "move from HubSpot to X" has already decided, they're just checking if the switch is survivable.

[George Maramigin @emaramigin]

@DaveTurewicz @n8n_io @prospeo_io @SlackHQ @smartlead_ai @attio @supabase @claudeai The separation point is usually workload, not company size. If the data needs separate permissions, high write volume, or its own compute, move it out. Otherwise, keeping the relationship model in Attio avoids another sync path.

[George Maramigin @emaramigin]

@DaveTurewicz @n8n_io @prospeo_io @SlackHQ @smartlead_ai @attio @supabase @claudeai One correction from doing migrations at the $10M mark: Attio doesn't stay a $1M tool. Once you add custom objects and record-referenced formulas, it does a chunk of what you're describing Supabase-beside-CRM for, before you actually need the separate database.

[George Maramigin @emaramigin]

@shubhambidgar16 Difference shows up clearest at handoff. "Called, left voicemail" is activity. "Pushed back on price twice, still engaged, wants ROI numbers before next call" is context. Most CRMs only force you to log the first one.

[George Maramigin @emaramigin]

@Syntaxia_ Most of that 57% comes from matching on company name instead of domain. Two reps create "Acme" and "Acme Inc" from a form fill and a manual add, and nothing flags it until someone runs a pipeline report and the total looks fine but the count is doubled.

[George Maramigin @emaramigin]

@sparqio The exit logic gap is the same thing that kills CRM data hygiene: everyone wires triggers for "when X happens, do Y" and nobody wires "when X happens, stop everything else." A single customer state field the whole stack reads before firing fixes most of this cheaper than a CDP.

[George Maramigin @emaramigin]

@stablewinston That's a task-recurrence problem, not a memory problem. In Attio you'd set a follow-up date on the record when you send the update, and it auto-resurfaces in your task list on that date whether or not they reply. No assistant needed, just a field that nags you back.

[George Maramigin @emaramigin]

@SilverHamster True for invoicing, but for a pipeline-driven B2B business it's often the CRM after all, because that's where the next 90 days of revenue lives, not just the last 30. Different failure mode: not "money stuck" but "money you don't know is coming."

[George Maramigin @emaramigin]

@digitalxroyalty Usually not the CRM itself, it's who built the fields. Let reps define stage criteria and required fields with sales, not just import a template from the last tool, and adoption resistance mostly disappears.

[George Maramigin @emaramigin]

@hackernoon The gap usually shows up at the deal-to-invoice stage: HubSpot closes a deal on signature date, Xero recognizes revenue on invoice or payment date. If those two ledgers aren't reconciled by deal ID, attribution reports quietly drift from actual revenue within a quarter.

[George Maramigin @emaramigin]

@hackernoon The one that bites most is "relationships" during dual-write windows: if old and new systems both accept writes even for an hour, you get orphaned child records nobody notices until reporting breaks weeks later.

[George Maramigin @emaramigin]

@cmyharish The part that usually kills it isn't the data provider, it's matching. If the field's missing, so is a clean key (domain, LinkedIn URL) to match records against, so the enrichment either mismatches or fails silently.

[George Maramigin @emaramigin]

@mattsonnen_pfi @FAMagazine The bloat usually isn't the feature set, it's required fields at the point of data entry. Every "just in case" mandatory field at deal creation is a rep who stops updating the deal. Cut required fields to 3-4 and adoption fixes itself faster than any redesign.

[George Maramigin @emaramigin]

@superpobe @attio Start with a Feedback object linked to the relevant person and company. Add only source, theme, urgency, and review status at first. More fields usually create noise before they create insight.

[George Maramigin @emaramigin]

@AleksAtanasov Step 3 is where most of this quietly dies. I've seen ICP firmographic fields get defined once, never revisited, so six months in half the "enriched" records are stale and the scoring model is scoring noise.