Would you fly in a passenger airplane with multiple, disconnected pilots? One controlling the engines, one for heading, one in charge of fuel, etc. Seems absurd to think that a plane could actually fly this way. But isn’t that analogous how we often set up and run our revenue organizations?
We have a pilot for marketing, one for sales, one for sales operations, and so on. These ‘pilots measure progress separately and differently, often incompatible and often opaque. As a result we are swimming in data but starved for actionable information that allows us to set predictions and then quickly and regularly course correct to achieve our revenue goals.
We regularly see three structural complications that are preventing startups and turnaround from being able to plan and predict growth more accurately:
- Sales and marketing silos are hardwired in place by similarly siloed applications
- Startups and turnarounds don’t have good enough data for forecasting and analytics
- Forecasting is done by extrapolating past data instead of using “adaptive planning”
For item 3 we’ll describe a system for startups and turnarounds to predictively plan and deliver revenue growth since analytics approaches that work for larger organizations break down here.
Mind the Gap – or more – the “MQL-SQL Iron Curtain”
One complication arises from the fact that startup and turnaround companies are ill-served by the sales- and martech applications that hardwire companies into a set of rigid, and fairly incompatible processes. Why do we have, say, Marketo or HubSpot for marketing and Salesforce or Dynamics for sales, and many of those apps are still poorly integrated? Why not integrate the two halves in one app?
Preferably one that can be used by the C-Suite directly, like if a CRO simply wants to know which marketing and sales investments have the best yield? Or how can they double down on those investments and reduce the poorly yielding investments? Instead they need an army of data and revenue operations high priests to tell them, often with too much of a delay.
Any startup or turnaround that’s trying to create a direct data-line of sight between first touch leads at the beginning of their pipeline, and closed won revenue at the end knows how hard it is, in the absence of complete, good and relevant data, to tie every lead source all the way through to revenue through a well-functioning lead-to-sales integration between applications. Never mind the technical challenges of integrating independent applications – even if you could, it’s impossible to do unless you have several quarters of good pipeline generation data available, period. What startup or turnaround has that?
Watching the “Analytics Theater”
Which brings us to the second, structural complication: Most forecasts predict the future by extrapolating from the past (for more detail on this topic, please see our earlier blog). What if you are a start up and don’t have much or any old data? Or what if your old data is poor or incomplete?
If there is no past pipeline data, no amount of AI-powered analytics with multiple Ph.D.-level data scientists on staff will give you meaningful answers. For later stage companies with good data and process hygiene, this makes eminent sense, and apps like Funnelcast or Bizible (now aka “Adobe Marketo Measure”) can provide deep insights into your pipeline and its drivers. But they need months or quarters worth of data to offer highly predictive conclusions.
Without good data, analytics and forecasting is futile. It’s really just a show, if you will, ‘Analytics Theater”. Great for the board, but not suitable to set challenging goals or help drive the daily behaviors of all those pilots you have hired.
Driving in the rearview mirror
Third, by bifurcating lead flows into their marketing and sales halves and then relying on absent, incomplete or irrelevant data, startups’ and turnarounds’ predictive revenue forecasting is further made difficult by betting on the wrong mathematical technique to do their forecasting with: Forecasting by extrapolating from past data (i.e. how the vast majority of forecasts are done). By definition this won’t work if lead flows are not continuous (how do you forecast across the MQL-SQL Iron Curtain if data from the apps on one side of the divide doesn’t make it to the apps on the other side?) and/or if the underlying sales and/or marketing data is of low quality.
So, if forecasting using statistical algorithms that essentially extrapolate into the future from the past is futile, then what works? One solution is to borrow from control theory, which has solved the problem of how to hit a future target without knowing any past data. It’s how a rocket makes it to the moon, it’s how a car navigates to a destination, and it’s how an airplane ascends to and maintains a desired altitude. These A to B journeys don’t analyze where they came from, they iterate into and optimize where they’re going.
Iterating forward into a target
Instead of looking at an unclear past to determine our future, let’s start with future goals, track our progress against them, and course correct. The scientific term for this approach is “indirect, multivariable, adaptive control”, which for this article we simply call “adaptive control”.
If we apply this approach to revenue planning and execution it means::
- Calculate backwards from a future revenue goal
- Select a target such as “we want to grow to $5M in net new revenue next year” or “we need to close 12 new deals this coming quarter”. This sets the forward direction for where we want to be revenue wise and by when, instead of trying to look back to where we came from. For example, If we’re shrinking, then simply extrapolating from past results will make the company crash.
- Then, architect a go-to-market system that generates new, first-touch leads from places where prospects can be found, and that can convert enough of those leads to achieve the target revenue growth.
- In the (temporary) absence of data, use a “reference model” to estimate growth
- A reference model is a flight simulator of sorts for your business that realistically models how your revenue will likely be generated (i.e. “how much revenue will come from which lead sources over what period of time?:).
- Over time, as real lead generation and pipeline progression data comes in, replace the underlying planning assumptions with actual, calibrated data.
- Perform frequent plan / actual comparisons to ensure you’re still on track
- Just like a plane in flight, compare the current flight path to the flight path we need to be on to make it to our destination.
- In a pipeline building context this requires monitoring both leading indicator KPIs (i.e. early funnel or activity metrics like no. of calls made or impressions generated, say) as well as lagging indicator KPIs (i.e. mid- and late-funnel outcome metrics like the number of SQLs or opportunities created or deals closed).
- Improve the predictive accuracy by periodically re-calibrating the reference model
- Continue to refine the model and assumptions. For example, update estimated conversion rates or lead velocities with more accurate coefficients.
- Institutionalize data and process hygiene as basis for a culture of accountability
- Break down the revenue generation process into measurable sub-steps to focus on solving specific areas of shortfall (e.g. Google ads are underperforming, or the Northeast region’s close rate is below average), vs. getting bogged down in unproductive political finger-pointing.
What gets measured improves
By implementing an adaptive control-based approach to revenue growth planning for B2B startups and turnarounds, the company leadership can track and improve a multitude of metrics from the get-go (vs. having to wait for 3 to 6 quarters until revenue operations has accumulated enough data to make reasonably accurate growth predictions that will satisfy the board). The kinds of the metrics that should be available from day zero are shown below:
- Lead conversion rates and velocities by lead sources
- Future revenue broken down by lead sources and who owns each source
- Detailed KPIs, i.e. by lead source show early-, mid- and late-funnel metrics
- Revenue and budgets broken down by lead source
Key to the ability to produce this level of detailed metrics from the very beginning is the reliance on adaptive control-derived reference pipeline model that forms the basis of the goals and KPIs shown here:
If armed with such precise targets, imagine how your company and colleagues can:
- Manage expectations:
- Handle board questions around the sources and justifications for your forecasts
- Management has an early warning system to enable them to course-correct in time and avoid getting blindsided over last-minute missed forecasts
- Drive cooperation and accountability:
- Marketing is accountable for and knows how to deliver high quality pipeline
- Sales doesn’t throw marketing under the bus, and instead picks up leads in time
- Both teams are generating growth on a scalable demand generation foundation
- Empower management:
- The CEO can drive revenue generation vs. being driven by end of quarter results
- The CFO and investors can approve revenue-optimized budget requests that are measurable and trackable
- The CRO has achievable revenue goals and can quantify what is needed from the rest of the organization, and also minimize the risk of over-forecasting
- The CMO can align their spend with revenue goals
- Revenue Operations has clear outcome and activity metrics to track
Results we’ve seen
By avoiding the three barriers to B2B startup and turnaround revenue success we mentioned at the beginning, i.e.:
- Disparate tech stack and processes that separate sales from marketing,
- Trying to extrapolate growth from bad, incomplete, or irrelevant data, and
- Switching from forecasting by extrapolating past data to forecasting using a forward-looking, adaptive and highly predictive planning methodology,
then revenue planning and delivery can transition from being something akin to a black art to a scientific, scalable and predictive growth planning and delivery system.
With precise KPIs and activity goals measurably aligned to revenue targets, this adaptive and predictive system will accelerate time to (higher) revenue, and reduce wasted budgets since the time spent in sub-optimal states is minimized. And, it lays the groundwork for an eventual, in-control, analytics-based demand gen system with excellent process and data hygiene.