I recently wrote about workplace anxiety and how data can be a great remedy. For this post, I’d like to expand on the topic of data, specifically how I’ve built and used “demand models”.
Demand generation models use various data inputs and then work backwards, using historical conversion rates and costs, to identify the budgets and scenarios you need to meet the demand in a given period.
A true demand model is tied to revenue, not leads.
It’s a bottom-up calculation where you start with your planned ending revenue number and you work your way up the ladder to figure out how much demand (in dollars) marketing needs to generate.
By distilling marketing and sales activities down to accessible numbers, your company leadership will have confidence that the goals are at least somewhat achievable, and what rates and values will need to hold true in order for the goals to be met.
The CMO can communicate more accurately with the CEO about how many actual opportunities and deals marketing will drive this quarter to support growth.
Steal our demand model template you can use to input your own data points and update regularly.
Here are the major data points I include in Metadata’s model:
Estimated start-of-quarter ARR – The amount of revenue you predict you’ll have at the start of the period you’re forecasting. For the sake of argument, let’s assume we’re measuring quarters.
Estimated end-of-quarter ARR – The revenue you want to get to by the end of the quarter you’re forecasting.
The net new ARR needed (the growth delta) – The difference between the start-of-quarter and end-of-quarter numbers, i.e. how much new ARR you need.
Net churn – Customers leave, new ones sign on. This number could be positive or negative. But it should be based on the actual $’s that are at stake for a given quarter, not just an average % for the year.
Most SaaS businesses sign on more people in the 2nd and 4th quarters than the 1st and 3rd, so straightlining an average won’t work.
Total ARR goal – The growth delta plus (or minus) net churn, giving you the total ARR goal for the quarter.
Let’s say that the total ARR goal number is $1.7 million. Marketing is not on the hook for sourcing and closing all of that in the quarter. The majority will come from existing sales pipeline you’ve been building up over the quarters.
Expected pipeline revenue – The next section of the demand model includes estimated revenue from current quarter and next quarter pipeline.
At Metadata, we have 6 opportunity stages, so I break this section down by each stage. Each stage includes the total revenue in that stage, the expected close rate for that stage, and then the expected revenue by multiplying these together.
I use historical close rates by opportunity stage to plug into the model and may either increase or decrease based on our current trajectory.
Total ARR to Source and Close in the period
Add together the expected closed/won pipeline revenue from the existing quarter and next quarter and this is the revenue you’re forecasting will close from existing opportunities. Subtract this number from the Total ARR Goal above and you have the amount of revenue that needs to be sourced and closed between now and the end of the period you’re forecasting.
We now start to work this revenue back to the activities we need to drive in Marketing and Sales to get to that number.
Average ARR selling price – In order to turn the revenue you need into the total number of new customers, you need to divide the total ARR goal above by the average ARR of each new customer.
Conversion rates – I then use stage-to-stage historical conversion rates to work the total number of new deals back to the number of early-stage opportunities I need to deliver. I then work that back into the number of demo requests I need to drive through Marketing.
How many demo requests then? I need to know my demo request to opportunity conversion rate, which will tell me how many demo requests I need to drive to turn into the early stage opportunities that will turn into closed/won deals.
Cost per demo request – As long as I understand my average cost to drive a demo request, I can now figure out how much marketing budget I need to drive the # of demos needed. If there is a delta (meaning I can’t afford all of them), then I get into a war room and start figuring out how to get additional lift on conversion rates, lower cost per demo, or places to find earned demand.
Note: I’ve included an ungated demand model template you can use to input your own data points and update regularly.
The one thing we haven’t discussed are sales cycles.
Sales cycles can be 30 days, 90 days, 120 days or more. This means the marketing I do today will likely not turn into company revenue for a quarter or more.
Metadata’s sales cycle is short enough that we usually have the pipeline/churn/revenue data within a quarter to feed a demand model. However, as we move upmarket our sales cycles will extend.
If your cycle is 90 days or less, then you’re within a quarter and don’t need to worry too much about this.
If it’s longer than that, you’ll want to do an analysis of how many deals are sourced and close within whatever timeframe you’re working from and use that to adjust the number of demo requests you need to drive so that enough of them close within that time frame.
For a basic demand model, these data points are a good start:
For a quarter or two, your model may only include 10 data points.
It’s fine in the beginning to just use estimates for churn and pipeline and then feed in more accurate inputs as you generate more sales data. Before long you’ll have 40 or 50 data inputs.
I do not recommend updating the demand model every day. One day it will say you need 100 demo requests, the next day it will be 85. Daily incremental data changes are not worth obsessing over.
So stick to doing updates once or twice a month. If there are discrepancies in your data, bake those into the next iteration and keep evolving. If your model says you have more than enough budget to pay for the demand this quarter, that’s great!
But don’t rest on your laurels. Spend the downtime finding that next round of demand for your product.
Whether your model delivers good news or bad about the demand/budget ratio, it serves a supremely important purpose: it reduces the dread of uncertainty, allowing you to plan ahead and make the best marketing decisions for your business.