If you’re in B2B and responsible for Go-To-Market (GTM) results, you already know: reporting is either a superpower or a total timesuck. Out-of-the-box dashboards rarely fit how your team works. And most “AI dashboards” are more about sizzle than steak. This is for people who want to actually track the right B2B metrics, in a way your sales, marketing, and ops folks will use.
This guide walks you through building custom dashboards in Einstein CoPilot. We’ll cover what works, what doesn’t, and how to avoid wasting your time on fancy but useless charts.
Why Custom Dashboards Matter for B2B GTM
The default dashboards in any CRM are generic. They’re made to impress execs who want pretty graphs, not operators who need to see, say, “Which channels are actually bringing in qualified pipeline?” or “Where are deals stalling?”
Here’s what you really want: - Metrics that map to your actual sales model. Not just “Leads created” but “Qualified pipeline by source,” or “Deal velocity by segment.” - Dashboards people look at. If your team ignores it, it’s just shelfware. - Fast answers, not more questions. You want to know what’s working, not get lost in noise.
Einstein CoPilot has some decent AI features, but don’t get distracted by automation for its own sake. The magic is in setting up the right metrics—then letting it do the heavy lifting.
Step 1: Pick Your GTM Metrics (and Ignore the Fluff)
Before you even open Einstein CoPilot, get crystal clear about what you want to track. Too many dashboards turn into junk drawers of every possible metric. That’s a fast track to nobody caring.
For most B2B GTM teams, start with: - Qualified pipeline by stage - Deal velocity (time in each stage) - Win rates by segment or channel - Lead-to-opportunity conversion - Source of closed/won deals - Sales cycle length - Average deal size
Pro tip: Don’t build a dashboard for every stakeholder. Focus on the metrics that actually move the business. If you’re not sure, ask: “What do we actually change if this number goes up or down?”
Step 2: Get Your Data House in Order
Garbage in, garbage out. Einstein CoPilot can only show you what’s already in Salesforce. If your data is a mess, fix that first—or at least be honest about gaps.
Check these before building: - Are opportunity stages and close dates being updated consistently? - Is lead source tracked reliably? - Are custom fields (like segment or product line) actually filled out? - Is your Salesforce data at least mostly trusted by sales and marketing?
If not, don’t waste time on fancy dashboards. Clean up the basics first.
Step 3: Set Up Your Einstein CoPilot Workspace
Now, log into Einstein CoPilot. Let’s not pretend setup is instant—you’ll need at least edit permissions and some patience.
Do this: 1. Navigate to Analytics Studio inside Salesforce. 2. Create a new app or workspace for your GTM dashboards—don’t clutter up existing ones. 3. Connect your data sources. By default, you’ll see Salesforce objects (Leads, Opportunities, etc). If you need Pardot/Marketing Cloud or external data, set those up now. 4. Decide on dashboard layout: Who’s going to use this? Sales managers? Execs? Make it easy for them, not just pretty.
Honest take: The UI is a bit clunky, and “AI suggestions” can be hit-or-miss. Stick to manual setup unless it’s really nailing your use case.
Step 4: Build Your First Dashboard (The Right Way)
Here’s where people mess up: they add every metric and chart under the sun. Resist that urge. Start with the core questions you picked above.
Let’s walk through a simple, useful dashboard:
A. Add Core Metrics
- Pipeline by Stage: Use a bar chart to show value of pipeline in each stage. Filter for “This Quarter.”
- Deal Velocity: Use a summary table showing average days in each stage.
- Win Rate: Pie chart or gauge showing percent of closed/won vs closed/lost.
- Lead Conversion: Line chart of leads turning to opportunities over time.
B. Make It Interactive
- Add filters for Sales Rep, Segment, Date Range. Let managers drill down without needing a new dashboard every time.
C. Give Context, Not Just Numbers
- Add text widgets to explain what each chart means.
- Include targets or benchmarks where possible (e.g., “Goal: 20% win rate”).
What to ignore:
- Anything that’s just “nice to know” but doesn’t inform action.
- Vanity metrics (raw lead count, random web traffic, etc).
- Overly complex AI predictions—unless you can explain them to a human in under a minute.
Step 5: Refine, Test, and Actually Use It
Once your dashboard is live, don’t just email it around and call it a day.
What works: - Sit down with a couple users (sales managers, ops folks) and get their honest take. - Ask which charts they actually use. Kill anything ignored for two weeks. - Schedule real reviews—if nobody brings up the dashboard in pipeline meetings, it’s not doing its job.
What doesn’t: - Building for every corner case. If someone wants a one-off deep dive, build a separate dashboard—not a Frankensteined monster everyone ignores. - Relying on “AI insights” to find problems for you. Often, these are just surface-level correlations.
Step 6: Automate (But Only What’s Useful)
Einstein CoPilot will try to recommend automations and AI-driven alerts. Some are handy, others just create noise.
Worth automating: - Email or Slack alerts for pipeline drops, stuck deals, or big changes in win rate. - Scheduled snapshots (weekly/monthly) to track trends over time.
Skip or be careful with: - Automated “insights” that aren’t actionable (“Lead volume up 3%!”—so what?). - Too-frequent alerts, which people just start to ignore.
Pro tip: One or two well-targeted alerts are more useful than a daily feed of trivia.
Step 7: Share, Train, and Keep It Simple
- Roll out slowly. Show the dashboard to a small group before blasting it org-wide.
- Train people on how to use filters and drill-downs. Assume nothing—most people only know basic Salesforce views.
- Document what each metric means. Avoid the “wait, what does this number include?” confusion.
What to Watch Out For
- Overcomplicating: The more metrics you add, the less anyone pays attention.
- Data trust: If your team doesn’t trust the inputs, the dashboard loses credibility fast.
- AI hype: Einstein CoPilot’s “smart” features sound fancy, but often need tuning. Don’t expect magic.
Final Thoughts: Keep It Simple and Iterate
A good dashboard tells you what’s working and what’s not—quickly. Start with the basics, see what people actually use, and add from there. Don’t let “AI” or “advanced visualization” distract you from the fact that most GTM teams just need clear, accurate, up-to-date numbers.
You can always make it fancier later. For now, focus on making something your team will actually use—and ignore the rest.