Every upsell starts with an opened opportunity
Why this matters. There are several ways to close an upsell, and closing is real work. But whichever model you choose, closers need a steady flow of opened opportunities to work. Opening is the step that’s easiest to leave unbuilt, and it sets your expansion number before any closer gets involved.
CSM closes
The CSM who owns the relationship carries the upsell. Cheapest model, but commercial tension can strain the trusted-advisor seat.
Dedicated upsell AE
A specialist rep works the installed base full-time. Clear accountability and a real outbound motion, at the cost of a quota-carrying head.
Original AE keeps it
The rep who closed the new logo keeps the account. Continuity is strong, but attention drifts to new pipeline when the month gets tight.
The rule: pick a closing model and move on. What determines your expansion number is how many opportunities get opened each month, and from which channels. That’s what the rest of this page builds.
Build the number from the touches you already make, and every miss points to the channel that caused it.
Where a top-down percentage breaks down
A common way to plan expansion is to forecast upsell revenue as a percentage of the customer base. It looks like a model, but it has no inputs you can act on: it’s an estimate rather than a calculation.
A top-down estimate
“We think we’ll get to 20% adoption by year end.”
A top-down percentage has no inputs, no owner, and no lever to pull when you’re behind. If it misses, there’s nothing underneath to inspect or fix.
- Where does the 20% come from?
- No stated basis
- Who owns hitting it?
- Not assigned
- We’re behind. What do we change?
- No lever
- What did we learn from the miss?
- No signal
The bottom-up model
Build the forecast from the ground up, using every touch your CX org already makes.
TouchesEach touch type has a monthly volume you can count.×conversionAnd a conversion-to-opportunity rate you can measure.= opened opportunities.
Opened opportunities × win rate × deal size = expansion revenue. Every number has an owner and a lever.
Every customer touch can open an opportunity
Your CX org touches customers every day: implementations, support tickets, CSM check-ins, outbound calls, email cadences. Each one is a chance to open an opportunity. The model is those channels, counted honestly, with a conversion rate on each. It has two layers: the human-to-human foundation, and the leverage channels built on top of it.
Layer 1: the human-to-human foundation
Volume × conversion = opps / mo.
| Human channel | Volume question | Conversion question |
|---|---|---|
| How many implementations go live each month? | How many surface a moment to sell: a new team, module, or use case? | |
| How many tickets come in each month? | How many reveal a need the next tier or product solves, and get flagged? | |
| What’s the cadence (weekly, monthly, quarterly), and how many touches does that total? | How many check-ins and QBRs end with a qualified expansion conversation? | |
| What does the outbound calling motion into the base look like: dials and connects per month? | What’s the connect-to-opportunity rate? | |
| How many customers enter an upsell sequence each month? | What’s the sequence’s conversion-to-opportunity rate? |
Layer 2: the leverage channels
Built on top of the foundation, not instead of it.
| Leverage channel | Volume question | Conversion question |
|---|---|---|
| What campaigns run into the base (webinars, newsletters, launches), and how many engage? | How many engaged accounts convert to an opportunity? | |
| How many usage signals fire monthly: limit hits, locked-feature views, upgrade clicks? | How many signals become a routed, worked opportunity? |
Every channel produces opps per month. Σ all channels = opened expansion pipeline.
Why the layers matter. Customer marketing and in-product funnels are well-documented motions with plenty of tooling: use them. But they amplify the human-to-human connections your team makes every day; they don’t replace them. A funnel email lands differently when a CSM was in the account last week. Build the foundation first, then layer leverage on top.
Build the forecast from the ground up
Put a monthly volume and a conversion rate on every channel. Replace the sample numbers with yours; the totals update as you type. When the forecast misses, you’ll see which row missed, who owns it, and which of the two inputs to fix: more touches, or better conversion.
| Channel | Touches a month | Conversion to opportunity (%) | Opportunities a month |
|---|---|---|---|
| ImplementationsGo-lives per month | … | ||
| Support ticketsTickets per month | … | ||
| CSM touchesCheck-ins and QBRs per month | … | ||
| Upsell AE outboundConnects per month | … | ||
| Email cadenceCustomers entering sequences per month | … | ||
| Customer marketingEngaged accounts per month | … | ||
| In-product funnelUpgrade signals per month | … | ||
| Opportunities opened a month | … | ||
Opportunities × win rate × upsell ACV. Sample inputs are illustrative.
Monthly figure × 12
First time through. You won’t know your conversion rates yet, and that’s expected. Estimate conservatively, instrument every channel, and replace the estimates with actuals as the data comes in. A ground-up model with rough inputs is still more useful than a top-down percentage, because every miss points to the input that was off. The sample numbers above are illustrative placeholders, not benchmarks.
Run it on a cadence
The model is only as good as the instrumentation behind it. Four habits keep it honest.
Instrument the source
Every expansion opportunity gets a required source field: implementation, ticket, CSM touch, outbound, email, marketing, or in-product. An opportunity without a source doesn’t get logged.
Count the opens
Review opportunities opened by channel against the model. Behind on volume? That’s an activity problem you can address this week.
True-up conversion
Compare actual conversion rates to the model’s assumptions, channel by channel, and update the assumptions to match.
Re-forecast from actuals
Rebuild the expansion forecast from trailing actuals. That figure is what goes into the plan.
How to roll this out
- Pick your closing model (CSM, dedicated upsell AE, or original AE), write down who owns opened opportunities, and leave the choice settled.
- List every touch type your CX org makes in a month and put a real monthly volume on each, counted from your systems rather than estimated.
- Add the required source field to expansion opportunities in your CRM before you do anything else.
- Train implementation and support teams on what a sell-to moment looks like, and give them a one-click way to flag it. Their job is to open opportunities for a closer to work.
- Set the model’s first conversion rates conservatively, then replace every estimate with a trailing actual as measurement data accumulates.
- Build one dashboard: opps opened by channel × month, with model vs. actual. This is the first slide of every expansion review.
- Retire the “% of base” line from your plan. If someone asks for the expansion number, show them the channel build-up behind it.