Monday Oct 25, 2010 by David Skok - General Partner, Matrix Partners
Great companies are almost always run by great management teams. And great management teams know that the only way to improve a process is to start by measuring it. Good metrics should also be actionable, and drive successful behavior. In this post I hope to help show how to figure out which metrics matter the most, and how to design them in such a way as to drive behavior that will lead to the results that you want.
This post is applicable to any kind of business. In a follow up post, I will use this technique to walk through the design of a set of metrics for a SaaS company. Since SaaS businesses (or any other subscription-based business) are different from standard software businesses, there are some interesting elements that we will uncover.
One way to look at how companies work is to imagine them as a machine that has Outputs, and Levers that you, the management team, can pull to affect it’s behavior.
Weak management teams have only a limited understanding of how their machines work, and what levers are available to affect performance. The better the management team, the better they will understand how that machine works, and how they can optimize its performance (what levers they can pull).
When we look to design metrics, we are looking to deepen our understanding of the machinery, and how it works. Well designed metrics will automatically drive behavior to optimize output from the machine.
Here is an example of a bad board meeting, which happens far more frequently than you might imagine. The company has just missed its quarterly revenue forecast. Good board members want to know two things:
As they ask management what happened, a common answer will be that the market was really tough, and deals just didn’t close the way that they hoped. They also don’t have a great plan for what they are going to do differently next quarter, other than hope that the market improves, and that more deals will close. There is a great saying for situations like this: Hope is not a strategy.
The better management teams answer those questions differently. They will gradually peel back the covers of the machine, like peeling the layers of an onion, and expose the true nature of the problem, which of course will also highlight what levers need to be pulled to fix the problem. Lets take an example, and look at how they might do this:
The contrast between the two approaches is stark. In the second case, it is clear that management will know how to fix the problem (by adding new traffic generation programs). They also know precisely how much additional traffic will need to be generated to reach the growth targets, and how many sales people are needed at a given productivity level, etc. etc.
What is surprising is just how few management teams really have their act in order in this area. For Web and SaaS businesses with smaller transactions at higher volumes, this kind of modeling and tracking is much easier, as web-based lead generation and marketing have easy to implement measurements, and the greater the volume of transactions, the more clearly patterns emerge. This is a little harder to do for channel sales, but still extremely valuable. And a little harder than that for direct sales situations with large deal sizes.
The secret to successful design of metrics is to start with the end goal and work backwards. In most companies, the end goals that matter the most are:
(You may wonder why we don’t have Revenue in this list, but read further, and and it will soon become clear.)
Let’s take the first of these, Profitability, and work backwards. Working backwards means looking at the components that make up Profitability:
Profits (EBITDA) = Revenue – Cost of Goods Sold – Expenses
So to focus the management team on driving profitability, we should also track and measure Revenue, CoGS, and Expenses. Obvious, isn’t it? Well the good news is that this same principle can be applied over and over again focusing on the components of Revenue, CoGS, and Expenses where needed.
So the next step is to take Revenue, CoGS, and Expenses, and break them down to the key components. Bookings is the pre-cursor to Revenue. So let’s look at Bookings as an example:
Bookings =No of deals closed * Average Deal Size
For Reseller Channels, we might be looking at something different like this:
Revenue = No of productive resellers * average productivity per reseller
(Note: in many businesses there are several categories of deals. e.g. there could be large deals, and smaller deals. Or their could be deals from two or more different categories of customers. So the formula may have more elements to it than shown above.)
Peeling back another level, we might find the following:
No of deals closed = No of productive sales people * Average Productivity per Sales person
There will also likely be another formula to compute this, which will look like the following:
No of deals closed = No of Trials * Average Conversion Rate
These two formulae clearly indicate some of the levers that we have available to increase Bookings. We can grow the number of trials, or grow the number productive sales people, or we could try to increase the average productivity of our sales people. However we need to make sure that we grow them both together, otherwise we could end up out of balance, and have too many sales people and not enough trials to feed them, or too many trials and not enough productive sales people to close them.
The next step would be to peel back the onion a few more layers:
No of trials = No of visitors to the web site * Average Conversion Rate to Trials
No of Visitors to the web site = Normal traffic + for each traffic generation campaign: target audience of each campaign * Conversion Rate to visitors
Each time we peel back a layer to expose the components, we gain a better understanding of our machine and the levers that we can pull to make it work better. For example in the above two formulae, we can see that a big driver of the model is visitors to the web site. But this can be expensive to increase. So the other variable that we can try to increase is the conversion rate for each campaign, and the conversion rate to trials. We can try to do this by altering campaign messaging and landing pages and using A/B testing to find the optimum creative content.
We might also decide to focus our efforts on increasing the average deal size. We could do this in several ways:
As with many good ideas in business, all of the ideas above are obvious, and follow common sense. However, you would be shocked to discover how rare it is to actually see businesses that have fully peeled back the onion to expose all the major variables and levers, and then implemented appropriate metrics to track these over time.
For every major variable that matters in our model, we will want to track how this varies over time. This will show us if we are succeeding in our efforts to improve things, and also give us early warning signs of any negative trends.
For most stages in a sales and marketing pipeline, we will want to track two metrics: how many prospects we put through that stage, and how effective were we at converting them to the next stage. For example:
|Stage in the Sales Funnel||No of Prospects||Conversion Rate|
|Campaigns to drive traffic||Eyeballs seeing the campaign||Conversion % to Visitors|
|Visitors||Site Visitors||Conversion % to Trials|
|Trials||No of Trials||Conversion % to Closed Deals|
|Overall Sales Process (start to finish)||No of Visitors||Conversion % to Closed Deals|
Another area where metrics can be extremely useful is in managing an inside sales (telesales) organization. Starting with the overall sales number achieved by the whole group, let’s peel this back layer by layer, to see what we can learn:
Not surprisingly we need to look at how each individual has done relative to the average levels to understand the strong performers, and the weak performers.
For the weak performers, it is likely that the number of deals closed will be lower than we want. The question is why? So what are the components that make up the number of deals that an individual closes? Assuming a sales process where each inside sales person is handed a queue of marketing qualified leads, and then calls these to try to schedule a demo, and the post the demo tries to close a sale, the components will be:
The above may not mirror your inside sales process, but hopefully the method of working backwards from the end goal, and peeling back the layers to expose the components will enable you to map out the metrics that matter to you.
We will also want to look at some metrics that cover the entire sales and marketing funnel from top to bottom. Here are some example metrics that are important at this overall level:
Lead source effectiveness:
Some categories like Expenses are made up of many line items, and we very likely don’t want to bother with metrics for every line item, we need to answer the question: How deep should we go with our analysis? The answer to this is pretty much common sense:
There is nothing in this article that should be surprising or earth shattering. It is all obvious. However, as is often the case in business, it is really easy to have the vision of what to do, but far harder to execute on that vision. In my experience the mark of a really well run business is that they actually have the systems in place to automatically produce these metrics. And they use those metrics as part of the management process to run the business.
One of the greatest things about putting in place the right metrics is that showing them to people will automatically change their behavior to try to improve the metrics. Furthermore, the metrics make it clear what levers they can use to change performance.
Working backwards from a specific Revenue target, management will be able to understand all the other elements that have to be put in place to reach that target. For example, if you want to hit $xm in bookings for the quarter, you can work out:
If you are in a channel model, you can work out how many productive resellers are required, and given a known conversion rate from newly signed resellers going through an on-boarding process, you will be able to work out how many new resellers are required, and how many on-boarding sales training sessions need to be run. Etc.
I have had the very good fortune to work with some excellent management teams that have helped teach me these lessons. In particular, I would like to thank the teams at HubSpot and JBoss who were very advanced in their use of metrics.
Watch out for my follow up blog post on SaaS Metrics and Levers to see what happens when I drill down on the key metrics for a SaaS business. (This is applicable to other subscription businesses such as Open Source.)
Skok is a General Partner with Matrix
Partners in Waltham, Massachusetts. You can find this post, as well as
additional content on his blog called For Entrepreneurs. You can also follow David on Twitter (@BostonVC) by clicking here.