A financial model with a balance sheet and cash flow is the most critical tool for every brand leader.
Here’s why. The only thing that will kill a brand is running out of cash. So self-preservation is reason one. Reason two is brand leaders must understand how their business works. They must understand the levers in the business, the relative importance of each lever, their ability to push those levers and the compounding effects of those levers over time. To put it bluntly, the proof they really understand their business is their understanding of their model.
A well constructed model provides all that.
In Part 1, I explain why this is critical for brand leaders so if you missed it, Part 1 is a 2 minute read.
In this post, I want to review the Do’s and Don’ts of brand models. These are 4 holy laws of modeling.
#1 Assumptions are levers you can control
The entire point of the model and its value is to show you potential visions of the future so if your model has variables and assumptions that you can’t influence what is it telling you? It’s like saying ‘and here’s where we win the lottery.’
So model assumptions must reflect the real levers in the business.
You must also prioritize the importance of these assumptions. Not all levers are equal. You have to decide which are most important. You may not know at the beginning so playing with the model will help you discover the relative impact of changes in assumptions which you can gut check.
#2 The model will change as the business changes
In the idea stage, you need a super simple model. You don’t really know what levers matter and you don’t know how well you can pull those levers. As your business grows, you will develop a much better sense of what really drives the business and how effective your team is at pushing the levers. Your model needs to reflect your increased understanding and competence.
Here’s a corollary to this:
Use the right model for your stage. If you are in the idea stage, don’t try to use your buddy’s 30 tab, multi-channel model for an 8 year old business.
#3 Models are a balance between precision, predictiveness and usability
It’s your responsibility as the leader i.e. the CEO or CFO, to set the right balance.
We have all seen models that are impressive in their detail. And model builders love to do this. It’s my biggest error in model building. Assumptions on assumptions, fancy features, ever more detail until there is only one person, the model builder, capable of understanding the model let alone using it. Increasingly important decisions are going to be made from the model, so the decision makers have to understand the model in order to trust it.
The model becomes so complex with so many variables and data feeds that simple questions can take hours or days to answer. In that case, the model has sacrificed usability for precision.
The other problem with precision is all those assumptions and tabs can create false precision. The model becomes precise, but wrong. It loses it’s predictive usefulness. So now, people can’t understand it, don’t trust it and have no idea what actions to take.
You as the brand leader have to strike the balance between precision, predictiveness and usability. And you will need to reassess this as your business grows (see law #2).
#4 Data is the master; not the model
A common problem for brands is every org and role has their own models with each variation built on different sources of truth and understanding of the business. Marketing is using a different model than inventory planning who uses a different model than ops and none match the one used by finance.
When this happens, there is a strong temptation to build the model to end all models. It can answer the 5 year strategy question and how much of SKU 12345 to buy next. It’s a direct input weekly cash model and an indirect monthly cashflow model. It’s cohort analyses combined with scenario plans. It has tabs for marketing and inventory planning and all the data and calcs are layered inside a massive file. If you have to download the model from a drive and it crashes your machine or you are seeing dozens and dozens of tabs, you are seeing an attempt at the model to end all models.
The real problem is two-fold. One, leadership has not conveyed to the team how the business really works. They have not established the levers and the weighting of those levers. In the absence of clear guidance, orgs fill the void with what they think is most important and their assumptions of how the business can or should work.
Second, data is siloed, inconsistent, not clean, spread unevenly. So even if everyone understands the levers, they may be using different data sets.
Technology is fixing this. Better data systems, AI and offloading compute from the model (e.g. client side to server side) is dramatically changing modeling. The ideal is common, cleaned and unified data being used by different modules that serve different functions and different purposes but all focusing on the right levers.
This is not unlike OKR’s or Rocks. Leadership has set the broad levers, but each org knows those levers break down into smaller pieces. Then every module and analysis is a sub-set of that top level understanding, just with more detail relative to the people needing it (i.e. marketing, ops) all using the same data.
#5 Start with the end in mind
I have found as I build models that I have a tendency to expand the scope as I build.
‘Oh, it would be cool if the model could do this. I can see wanting to tweak that…’
So, a new assumption is layered in. Or more likely, some complicated logic has to be built in to allow the model to operate under mode A or B. Implementing the new feature becomes a new puzzle to solve and puzzles are fun. Time is added to the build, but also complexity. I will come back to a model after a couple weeks or even a couple days and have to work my way through my own logic. If I am confusing myself, imagine what a new user experiences!
To prevent this, before I start modeling, I force myself to sketch out the analyses I want to see. I do this with pen and paper. This process always reveals tabs or logic I need to work out. Once I have this in place, I summarize the analyses and tabs and sections I need to build into a list. I now have a good idea of the model in my head and I just start checking items off the list. This helps me get to the end product much faster and provides a brake against the ‘cool new feature’ creep. I can always add that to the bottom of the list if I really want.
What if you are not the model builder, but are the end user and you are instructing someone on the model you need?
I still go through the process above. But then I open up a spreadsheet and I create the outline of the analyses and outputs I need. I just dump in dummy data i.e. lorem ipsum but numbers. And I will even add the formatting conventions I want to see if it’s someone I have not worked with before.
I then review these outputs with the modeler so they understand exactly what I need and it also helps them discover logic that may be required and to identify ways of doing things that are different from what they normally do.
As I gain experience with the modeler, I will ask them to produce these output outlines prior to beginning modeling so I can add tweaks or clarifications.
If you are struggling with this issue, I offer brand leaders a 10 day sprint to a clear plan for cash, capital, and profitability. Let’s talk.
NOTE: this post was originally posted Nov 4 and then updated Nov 24 to add the 5th Holy Law.