If you're tired of sales forecasts that feel more like hopeful guesses than grounded predictions, you're not alone. This guide is for sales managers, ops folks, and anyone who's ever had to explain a missed number. We'll walk through how to use Attention to actually turn your historical sales data into something that helps you plan, not just panic.
Why Historical Data Beats Gut Instinct
Let’s get something out of the way: “gut feel” sales forecasts are almost always wrong. Historical data, on the other hand, tells you what’s actually happened. It’s not perfect, but it beats wishful thinking every time. If you want to be taken seriously by your execs—or just sleep better at night—ground your forecasts in reality.
Step 1: Get Your Data in Order
You can’t forecast what you can’t see. Before you open up Attention, make sure you’ve got clean, usable historical data. Here’s what matters:
- Deal data: Closed dates, deal size, sales stage at close, rep assigned, etc.
- Pipeline data: Current open opportunities with projected close dates.
- Activity data (optional): Calls, emails, meetings—sometimes useful, but don’t stress if it’s messy.
Pro tip: Garbage in, garbage out. If your CRM is full of outdated or duplicate deals, take a day to clean up. You’ll thank yourself later.
What to Ignore
- Don’t waste time with vanity fields (e.g. “Deal color,” “Vibe score”). Stick to what actually moves the needle.
- Data from so long ago that your sales process or product was totally different? Leave it out.
Step 2: Connect Your CRM to Attention
Attention works best when plugged into your main CRM (Salesforce, HubSpot, etc.). Here’s how to get set up:
- Sign in to Attention.
- Go to the integrations page.
- Connect your CRM. You’ll need admin rights.
- Select the data you want to sync. Focus on closed-won, closed-lost, and open opportunities.
Attention’s integration process is pretty straightforward, but if you hit a permissions snag, ask your CRM admin for help. Don’t try to upload massive CSVs by hand unless you like pain.
Got legacy data? If some of your historical deals aren’t in your current CRM, import what you can. But don’t bend over backwards for ancient spreadsheets—use what’s reliable.
Step 3: Choose Your Forecasting Model
Attention offers a few different ways to generate forecasts. Here’s what you need to know:
- Simple weighted pipeline: Multiplies deal value by probability to close. Fast, but not nuanced.
- Historical trend analysis: Looks at seasonality, win rates, and past performance. Good for spotting patterns.
- AI-driven forecasting: Uses machine learning to find hidden trends. Can be useful—but only if your data’s solid.
Honest take: Start simple. If your data is limited or you’re new to forecasting, basic models are easier to sanity-check. The AI stuff can be impressive, but don’t trust it blindly—especially if your team or market is in flux.
Step 4: Set Up Your Forecast in Attention
Now for the main event. Here’s how to actually generate your forecast, step by step:
- Go to the Forecasts section.
- Pick your forecasting method. (See above)
- Set your time frame. Monthly or quarterly tends to work best.
- Choose your data set. Filter out outliers or weird deals that don’t reflect your current sales motion.
- Run the forecast. Attention will crunch the numbers and spit out a projection.
- Review the forecast confidence score. Pay attention to this—low confidence usually means your data is thin or erratic.
What to Watch Out For
- Outliers: One massive deal can throw off your numbers. Exclude anything that’s not repeatable.
- Seasonality: Attention can spot patterns, but you know your business best. If Q4 is always a blowout, sanity-check the forecast.
- Pipeline quality: If your open deals are mostly fluff, your forecast will be off. Gut-check the pipeline before you commit.
Step 5: Dig Into the Details
A good forecast isn’t just a number—it’s a breakdown you can act on. Use Attention’s reporting features to answer questions like:
- Which reps or segments are on track (or not)?
- Are certain products or regions dragging down your average?
- Where are deals getting stuck or delayed?
Pro tip: Share these details with your team. If everyone knows where the bottlenecks are, you can actually do something about them.
Step 6: Adjust and Re-run as Reality Changes
No forecast survives first contact with reality. That’s not pessimism—it’s just how sales works. The key is to update your forecast as new data comes in:
- Update your CRM weekly. Closed-won, new deals, deals pushed to next quarter, etc.
- Re-run the forecast in Attention. Set a cadence (weekly or bi-weekly is usually enough).
- Track forecast vs. actuals. If you’re consistently missing, figure out why. Was the pipeline padded? Are close dates fantasy?
When to Ignore the Forecast
If your company just launched a new product, entered a new market, or changed the sales motion, all bets are off. Historical data won’t help much—use your judgment and be upfront about uncertainty.
Step 7: Communicate and Document Assumptions
You’ll get the most out of your forecast if everyone knows what went into it. Before you share the number, write down:
- What data you included/excluded
- Any known pipeline risks (big deals with shaky close dates, etc.)
- Any manual adjustments you made
Pro tip: Make this part of your forecast report, not just a mental note. It’ll save you headaches when someone asks, “How did we get this number?”
What Works, What Doesn’t, and What to Ignore
What works: - Using recent, relevant data - Regularly updating your forecast as things change - Focusing on pipeline quality, not just quantity
What doesn’t: - Blindly trusting AI or black-box models - Including every single deal, no matter how old or weird - Setting and forgetting your forecast for the whole quarter
Ignore: - Overly complex models if your data is thin - Forecasting down to the dollar (round to the nearest $1k or $10k—it’s fine) - Forecasting for the sake of forecasting (use it to drive action, not fill in a slide deck)
Keep It Simple, Iterate Often
Don’t get hung up on making the “perfect” forecast. The goal is to get a realistic sense of where you’re headed so you can make smarter decisions. Start with what you’ve got, be transparent about the limits, and update as you go. Over time, you’ll get better—and your forecasts will actually mean something.