Hi,

It is fully fall now, with Halloween almost upon us (I'm going as Star Trek's Jean-Luc Picard after being mistaken for Patrick Stewart while in NYC last December), but my summer really ended on August 7th when I cut my knee while trail running. It turned out to be a pretty hard hit, and while the ER treated just the laceration I wasn’t able to bend my knee while it healed and I’m working to regain functionality. It’s hard for me to sit on the sidelines! I’ve taken it easy for a while, read more, and just plain let go of things I couldn’t do while literally lame. I certainly appreciate my typical functionality--being able to navigate an airport, walk through a grocery store, and even get out of bed easily. I certainly didn’t know what I had until it was gone. Now I'm working my way back with lots of PT and help from Nikki and Wendy at My Whole Body in Boulder. Highly recommended.

So from full appreciation of that, I hope this finds you healthy and well and enjoying our great world.

Meanwhile I've been working on a new website--see below for more about that.

Best regards,

David

I’m relaunching my websites, brodegroup.com and brodetelecom.com. I found the old material a bit bombastic, mainly because I relied on marketing folks who wanted me to punch above my weight as a solo practitioner and project the image of a larger firm. But I’m reconciled to reality, and I am a solo consultant. I think these new sites are more straightforward about what I do professionally. I’d welcome your feedback. I believe brodegroup.com is up but brodetelecom.com won't be for a few more weeks.

What’s the difference between a good model and a great model? I recently reviewed a model from a startup. The founders were engineers and had built a pretty good model. It had clear production volumes and prices. It had clear COGS for raw material inputs and process costs. Capex was very well specified. The cash flows looked technically correct and they calced IRRs. In all, it was a very solid model. But as I thought further I realized that there were big things missing from the model they’d need to show to investors. These are issues I see time and again. If you are looking to raise money, be sure your model can do these six things.

1. Have the Ability to Show Multiple Strategies.
This model, basically, showed projections for building one big factory using the company’s new technology. The only dial that could be turned in the model is to change the size of the factory. But sometimes investors want to try out different strategies. For example, what if we kept building factories after the first one proved itself? The model can’t handle that, and thus they don’t have the ability to show a huge upside case where we start replicating the template using low-cost financing.

2. Explain unit economics.
Even though the model shows this factory making tons of EBITDA over time, I have no idea what the marginal economics look like. I get that I need this really expensive factory, but what are my COGS? At what volume is the factory profitable at the margin and over a longer period? All the data about the detailed inputs and process engineering is in the model, it just isn’t summarized in an understandable way. As a result an investor will have no idea about the overall economics, e.g. IRR of the whole enterprise as if it were funded completely with equity.

3. Communicate clearly instead of being technically correct.
This category is more art than science, but I find that people can struggle to understand a model with extra complexity. In this case there were two issues. First, what I described as the factory earlier actually had three components and together they produced five products. There was much detail about the separate components, which is great, but my experience is that people need a simpler story. So if you talk about one output (which has a blended price) and a single simple “machine” that makes it, well, most people can follow that story.

One oddity in this monthly model was that it based production off the number of days in a month. While it’s true that October has fewer days than August, these technically correct variations in months added no value and instead just made numbers look more random. That doesn’t help investors who are trying to get comfortable with your plan.

4. Be able to show multiple financing scenarios.
This model contemplates one very specific (and non-standard) financing scheme. At times the financing assumptions were intertwined with operations calculations. It’s fine, of course, to show a financing view. But many investors want to suggest different structures, and a good model should be able to run different financing scenarios. In general, I find it’s best to be able to handle traditional funding structures (e.g. equity investment in return for shares in the company) and to clearly show splits and returns for all parties.

5. Show operational progress over time..
This model had a very simple view of the factory: once it opened, it just kept turning out the same number of widgets every day. My experience is that things don’t work that way, and it looks naive to suggest that in a model. I like to show some sense of a learning curve. For example, you could show more downtime in the first year and increase utilization rate over time.

6. Show sensitivities on commodity prices.
The factory in this model sold commodity products, and investors will know those prices. I was hoping for some analysis of the covariance of the prices of both inputs and outputs. With that we could determine a utilization rate at which the factory made economic sense.

So this wasn’t an obviously bad model, riddled with errors. In large part it seemed accurate. It just didn’t pop and present their case in the best way possible.