HubSpot, Marquee, Partner, Sales Management 2 Comments
3 Proven Methods for More Accurate Sales Forecasting
According to CSO Insights, 60% of forecasted deals do not actually close. Unsurprisingly, the data also shows that one in four companies are unhappy with their ability to accurately forecast sales. That’s a lot of revenue being left on the table and a huge opportunity that I personally, am not prepared to miss out on.
While I’ll admit it’s near impossible to predict the future, there are certainly more effective and accurate ways to forecast expected revenue than what I see many sales managers doing today.
In this post I’ll discuss three sales forecasting models that have proven to be effective for us at HubSpot. In fact, we’ve seen that a combination of all three has actually given us the most accurate predictions.
I’ll give a high level overview of the concept and forecasting methodology, but I also recommend you test and tweak them to fit within your own business model before rolling them out to your teams.
Let’s get started.
#1. The Lead-Driven Model
Concept: This model involves analyzing each of your lead sources and creating a forecast based on the value of each source.
Looking at the very beginning of the buyer’s journey can tell us a lot about how that journey will end. It’s like a bad romantic comedy; there are early tell tale signs of how the story will end if you’re in any way perceptive.
By assigning a value to each of your lead sources or types, you can get a better sense of the probability for each of those leads to turn into revenue.
For this model, you’ll need the following metrics:
- Leads per month for previous time period
- Lead to customer conversion rate by source
- Average sales price by source
Total No. of Leads
To calculate the total number of leads needed in a given time-frame, you need to look at the overall revenue goal and divide that number by your lead value.
Leads Needed = Desired Revenue / Average Lead Value
As a very simplified example, if your sales team needs to hit $100,000 in revenue for the month and your average lead value is $450, that means you’ll need to generate 223 leads in that time period.
100,000 / 450 = 223
Note, you should consult with your marketing team to learn what upcoming initiatives they have planned and where they expect lead flow to come from as lead values will vary, channel to channel.
To calculate the lead value per source you multiply the average sales price by the average close rate for that source.
Lead Value = Average Sales Price * Conversion Rate from lead to customer
For example, if I know my leads from paid advertising spend on average $2,000 with us and they convert at a rate of 10%, the lead value of each of those leads would be $200.
$2,000 x 10% = $200
Average sales price per lead
To get your average sales price by source you simply have to look at the data set for your entire customer database and bucket them by lead source. If your CRM doesn’t have this reporting functionality, you can export the data into an excel file and quickly get the average sales price from there.
Here’s a simplified version of the lead-driven model.
While this is a great starting point, there are other factors that can alter your end results which must also be considered:
The average sales cycle may vary for each lead type or source so you should conduct an extra layer of analysis on time to purchase and factor it into your forecast.
Other business initiatives might change your conversion rates such as improvements in the sales process, price changes or discounts, etc. Look at the value per lead for each source on a trailing 90 day period to stay current with other business changes.
Changes in the lead generation strategy – Marketing may adapt their plans based on learnings or evolving trends. It’s important to stay aligned with them to ensure your expected lead volume and conversion rates are accurate.
Not all lead sources are as easily identified as inbound leads are but it’s important to include all leads in your forecasting. If you can’t identify the source for a group of leads, you can bucket them as ‘other’ and include them in your forecast.
#2. The Opportunity Creation Driven Model
Concept: This model helps you predict which opportunities are more likely to close based on demographic and behavioral data.
If we go back to the movie analogy, it’s often easy to predict who is going to fall in love with who, who’s the hero, and who’s the villain based on what the characters look like and how they behave and interact with each other.
Predicting an opportunity’s likelihood to close is very much like this. By looking at demographic and behavioral data, we can get a better sense of the probability to close and the expected value of the deal.
In this model, we base our predictions on the internal factors at the prospect company that mostly correlate to closed business, e.g. company size, contact’s role, current technology in use, etc.
To illustrate, let me take you through the way we implement this model at HubSpot.
We have found that the simplest way to evaluate an opportunity’s potential to close is to look at the size of the business, (number of employees, annual revenue, etc).
However, there are many other factors that can determine the fate of an opportunity: the role of your contact within the decision making process, behavioral patterns, previous interactions with your company, etc. This second layer of analysis is called lead scoring and is something you can work with your Marketing team on to define and set up.
At HubSpot, we score our leads between 1-100, with 100 being the best fit. We take this numerical number and bucket it into A,B,C,D values for ease of use.
While these factors tend to correlate most with closed business at HubSpot, they could be completely different at your organization. The key is to look at the historical data for your customers; not just those that close but also those that retain and become referrals. These are the types of companies you want to prioritize.
Once you have your scoring system in place you can calculate the estimated value of each opportunity in your pipeline.
Expected Value of Opportunity = Average Sale Price * Close Rate
Below is a simplified forecast of expected value per opportunity based on lead score and company size, with an average sales price of $4,000. For this to work you need to know the close rates for each of your lead buckets.
*marked with green = primary focus
**marked with orange = secondary focus
What I love about this model is that it looks in-depth at the real potential of each opportunity which really helps my reps prioritize the opportunities with a higher expected value.
For this model to work you need to have a well-defined criteria for opportunity creation. However, even with this in place you’re relying on your sales reps to follow procedure and remain consistent in their administrative activities so you’ll need to keep an eye on it.
You have to build an opportunity scoring system or use a program that can automate the process, which can be time consuming and/or costly.
You need to be able to trust the data your opportunity scoring system uses to assign the score. I recommend testing the new system with one salesperson for a set amount of time before rolling it out to the greater team.
#3. The Opportunity Stages Driven Model
Concept: This model predicts the probability of an opportunity to close based on where the prospect currently is in your sales process.
This one is pretty logical. In sales, if you know the average sales cycle for your product and you have mapped out the stages involved for someone to get from early stage awareness to decision made, you can get a good sense for their likelihood to close within the current forecasting period.
Here’s an example of the deal stages you might use for your sales process and the probability associated with each one:
- Appointment Scheduled (20%)
- Qualified to Buy (40%)
- Presentation Delivered (60%)
- Contract Sent (90%)
- Closed Won (100% Won)
- Closed Lost (0% Lost)
In this model you create your forecast for expected revenue by multiplying the amount associated to each opportunity by each opportunity’s probability of closing.
Expected Revenue = Deal Amount * Probability to Close
For this model to work you will need a well-defined sales process with a detailed outline of the activities that need to happen in order to progress the deal forward towards closed won. Once you define your deal stages you then assign a probability to close for each one.
Below is a template you can use to map out your sales process. You can download an editable version here.
Here’s how this model should look:
For optimum accuracy, you’ll need a CRM system that allows you to assign the win probability for each stage in the sales cycle. Having a system in place will avoid any potential human error like manually inputting incorrect figures or forgetting to update it at all.
Getting buy-in for any system you adopt is incredibly important; if your team doesn’t use the system, you’re always going to be shooting in the dark.
It’s also a good idea to do a routine check every 6 months to see if your team’s performance is higher, lower or about the same as you anticipated when you initially set the probability. You should adjust the rates as your team becomes more productive and improves their conversion rate.
Old opportunities that have been sitting in your pipeline for months (maybe years) can affect the forecast. Make sure your data is fresh and the opportunities are updated regularly.
The probability factor is critical in this model so look at historical data and calculate it based on the performance of previous opportunities.
You need to have a very well-defined list of actions that need to happen before a deal can be moved into the next stage. Without clear guardrails over this part of the process you lose accuracy.
Sales Forecasting Using a CRM System
The table versions of these sales forecasting models are ideal when you’re just starting out. However, if your organization is more established, the best thing you can do is to customize the reporting section in your CRM.
I’d love to hear how these models work for your business or if you’ve used other methods for sales forecasting that have proven to be effective. Drop me a line in the comments below!
Also published on Medium.
This is a sponsored guest post from a Sales Hacker partner.