The 6 Attio configuration mistakes I fix every week
Short answer: The six Attio configuration mistakes that show up most often are: giving everyone Admin access, creating too many pipelines, using the wrong object for the job, not tracking who owns each data source, letting email sync auto-create unwanted records, and skipping "who updates this field" for every attribute. Each one is fixable in under an hour once you know where to look.
We set up and audit Attio workspaces for a living. Most workspaces that stall or go off the rails share the same handful of problems. None of them are hard to fix. All of them are easy to avoid.
Here are the six.
---
1. Giving everyone Admin access
The mistake: when setting up a new workspace, the path of least resistance is to invite teammates as Workspace Admins. Fewer permission prompts. Everyone can do everything. Setup feels fast.
What actually happens: any team member can delete objects, change workspace settings, create or remove attributes, and break automations. In a small workspace, this rarely causes an incident immediately. It almost always causes one eventually, usually when someone deletes a field that feeds a report nobody noticed was connected.
The fix: give Admin access to the workspace owner only. Everyone else gets Member access. Members can create and edit records, use automations, and configure their own views. They cannot change the object model or workspace settings.
If someone needs to run an import or change a pipeline configuration, they ask the admin. That five-minute friction is worth having.
---
2. Pipeline sprawl
The mistake: creating a new pipeline every time a process looks slightly different. A team that runs two products ends up with six pipelines. A team doing outbound and inbound ends up with four. Within three months, no one knows which pipeline is the right one and deals land in the wrong place.
What actually happens: reporting breaks across pipelines because there is no single view of pipeline health. Reps disagree about where to log a deal. Managers cannot run a meaningful funnel report.
The fix: one pipeline per distinct go-to-market motion. Not one per product. Not one per sales rep. Not one for "old deals" and one for "new deals."
A distinct motion means deals move through fundamentally different stages and are owned by different processes. Enterprise and SMB often qualify. Inbound and outbound rarely do; they share the same stages once the lead is qualified.
If you already have pipeline sprawl, audit each pipeline by last-modified date. Consolidate any pipeline where deals have not moved in 60 days.
---
3. Using the wrong object for the job
The mistake: this takes two forms. The first is using a List as a pipeline when the Deals object is the right tool. The second is creating a custom object for something that should be an attribute on an existing object.
Both come from not thinking through the data model before building.
What actually happens: Lists used as pipelines cannot report correctly, cannot trigger deal-based automations, and break when you try to connect them to other records. Custom objects created for what should be attributes bloat the workspace and create relationship complexity that nobody needs.
The fix: use the Deals object for anything with a monetary value, a stage lifecycle, and reporting requirements. Use Lists for views into records, not as a stand-in for deal tracking.
For custom objects: ask whether this thing has its own lifecycle and needs to be tracked independently over time. If yes, it is an object. If it is metadata about something else (a project phase, a contact's role, a contract status), it is an attribute.
When in doubt, model it as an attribute first. Attributes are easier to remove than objects.
---
4. No data source ownership
The mistake: multiple people import CSVs, connect email sync, plug in enrichment tools, and add integrations without anyone coordinating. Each connection creates records. Nobody tracks which records came from where.
What actually happens: duplicate records from multiple import passes. Unknown-quality data in key fields because three enrichment tools all wrote to the same attribute and the last one to write won. Reports that look inconsistent because the same company appears twice under different names.
The fix: designate one person as the workspace data owner. Every new data source, every import, every integration goes through that person. They maintain a short log of what data sources are active, what fields each one writes to, and what the dedup rule is for each one.
This is not about gatekeeping. It is about having one person who knows the answer to "where did this record come from."
For teams using email sync: configure it to create records only for people and companies already in the workspace. Do not let the sync create net-new records automatically. It floods the workspace with contacts from newsletters, tool vendors, and support threads.
---
5. Auto-created records polluting the workspace
The mistake: connecting email sync and calendar sync with default settings and not reviewing what gets auto-created.
What actually happens: Attio creates Company records from every email domain it sees. That means your project management tool, your accounting software, three personal Gmail accounts, every vendor, and every newsletter end up as Company records. Within a month, the workspace has hundreds of records that have nothing to do with customers or prospects.
This breaks list views, makes search noisy, and inflates record counts in reports.
The fix: go to Settings, then Email and Calendar Sync, and configure which records can be auto-created. The setting to change is which domains are excluded from auto-creation. Add your own company domain, common personal email domains, and any tool vendor domains immediately.
For workspaces that already have this problem: run a filtered list of Company records with no deals, no notes, no tasks, and no associated people who are active contacts. Review that list weekly and archive records that do not belong. This is a 30-minute task the first time and a 10-minute task each week after.
---
6. Skipping "who updates this field"
The mistake: creating an attribute without assigning ownership or a trigger. The field exists, looks good in the data model, but nobody updates it because nobody knows whose job that is.
What actually happens: six months after the workspace goes live, the field is 30 to 50 percent empty. The one automation that reads it breaks silently because it hits blank values. Reports that depend on it are wrong. The team stops trusting the data.
This is the most common cause of "our CRM doesn't work" complaints. The CRM works fine. The fields are just empty.
The fix: for every attribute you create, define two things before the attribute goes live.
First, who updates it. A specific role ("Account Executive"), not a vague category ("whoever handled the call"). For high-stakes fields, write this in the attribute description so it shows up as a tooltip in the record.
Second, what triggers the update. "Account Executive updates this field after every demo call" is a process. "Automation updates this from the deal-won trigger" is better because it is not optional.
If you cannot answer both questions for a field, the field is not ready to go live. Keep it in a draft object or a planning doc until the process exists.
---
How to audit your workspace right now
If you want to check your workspace against these six mistakes in the next 30 minutes:
- Settings > Members > check who has Admin role. Should be one or two people maximum.
- Deals > Pipelines > count how many active pipelines you have. More than three is worth reviewing.
- Records > check your custom objects. For each one, ask whether it has its own lifecycle and reporting need.
- Settings > Integrations > list every connected data source. Is there one person who owns this list?
- Settings > Email Sync > confirm excluded domains are configured.
- Pick three attributes in your most-used object. Open the attribute settings. Is there a description that says who updates it?
Most teams find two or three of the six in that 30-minute pass.
---
Want a second pair of eyes on your setup?
We run a free 48-hour Attio workspace audit. You add us as an Attio expert (no extra seat, no billing). We send back a one-page written teardown ranked by impact, the three highest-leverage fixes with the exact setting change, and a 5-minute Loom walking through the top fix. No call, no pitch. Five slots a week.
Need help with your Attio setup?
We migrate teams, build data models, wire automations, and train Claude agents inside your workspace. Discovery call is free.
Book a free discovery callReady when you are.
Two ways in. Pick the friction that fits.