At some point, and usually pretty early on in the life of a startup, a founder needs to put together a financial projection model. By this I mean they need to build an analysis of the expected costs and anticipated revenue of their budding business.
I usually suggest doing this as early on in the planning stages of the business, even before any other part of the business plan; because this can more easily help you differentiate a good business from a bad business, and will also allow you to understand your financial needs. As with the rest of the business plan, this is a forever changing document that must be consistently updated and reworked.
So how does one balance the need for detail and thoroughness that is essential to functional projections and the very real lack of information that will plague all and every startup?
I am not sure, but having built a few models already I can tell you that for each case it was different. For DVA (sold minor league advertising in the booklets and programs they hand out at the entrance) we had a one tab spreadsheet that outlined our printing costs, our distribution costs, and our overhead, and allowed us to plug in advertising pricing and understand our income from the business. It was really simple and it did the job. We probably couldn’t scale with it, and if we ever needed funding that document wouldn’t have been laughed out of the room, but it served its purpose.
Our current financial model for Emerging Demographics is much more complicated and intricate. It has several business lines, multiple websites driven by different assumptions, costs separated into over 40 very specific drivers, it is broken out into multiple locations, and we align our model to market trends. All in all, it is about a 20 something tab spreadsheet that that took an unfathomable time to build, but which we can update very easily and gives us a very thorough analysis of what our business looks like. Also, because it is so broken out, we are able to constantly update any assumption for which we get better information.
So which is better? I don’t really have a clearly defined opinion on this only that it depends on the business and the founder. I liked keeping the model simple the first time around, because it saved a lot of time that we were able to spend on other parts of the business, and most projections are completely wrong anyway. But building such a detailed model was incredibly insightful into the business and definitely made us smarter as to how the business needs to be structured.
If I was doing it again, I probably would go full out and build the projection model as detailed as possible, keeping in mind that a lot of it will be wrong, but using it as a tool for understanding my business.
Does anyone have any thoughts on this?
Popularity: 100% [?]
June 3rd, 2007 at 11:23 am
Yes, I have a career’s worth of thoughts on this, having worked on business plan templates since 1983 and stand-alone business plan software since 1994. Perhaps more important, I’ve had to support a financial model used by several hundred thousand people over more than 10 years. Support means maintaining and explaining, and, over time, productizing. It has to withstand a lot of pushes and pulls to be useful — meaning it reflects accrual accounting, payables, receivables, working capital in inventory, and cash flow — for a lot of different companies.
At the core, the problem and its solution revolve around a workable compromise between the power of detail and the problem of diminishing returns. We want the zoomable detail of accounting but we need increasing summary and aggregation as the timeframe heads further into the future.
That may sound conceptual but to make it concrete consider detail like depreciation. You can’t build useful projections without considering depreciation because it reduces income but not cash. The tempation, therefore, is to build a model that mimics the detail of accounting. And that doesn’t work because as you project further into the future it becomes increasingly less useful to attempt depreciation as a function of specific assets purchased and specific depreciation formulas. You should project depreciation as an estimated increase, not by guessing the assets. Specific example: if you estimated $25K for depreciation in the third year of the future, make it $35K in the fourth.
I have a post at blog.timberry.com called “Planning vs. Accounting” that gets at the underlying problem of confusing the two dimensions of financials. Accounting starts today and goes backward into time in ever-increasing detail. Planning starts today and goes forward into time in ever-increasing summary and aggregation.
I also have a detailed treatment of my financial model up at www.timberry.com/fm/ … it does a sales forecast, personnel plan, income statement, balance sheet, cash flow, and a start-the-balance sheet table that is either starting costs (for a start-up) or past performance (for an ongoing company). The model depends on some fundamental principles that give us the math in the background to manage cash flow, such as the magic of assets always being equal to capital and liabilities, so we can work some key projections backwords from manageable assumptions. And to make that work, we have to make some simplifying assumptions that require a summarized and aggregated balance sheet and cash flow, 30-some lines, not 300.
As you build that kind of a system, you end up defending a standardized model against a steady pressure for chipping away at it by adding detail. Of course it starts out with expandability of sales lines and expense lines and cost lines, but you do have to standardize the balance sheet to make the cash flow work.
The temptation to build more detail is always in conflict with the benefits of standardized, documented, and productized. And that’s where I get on a soapbox, because there is huge benefit to a working financial model that does reflect proper financial workings and does predict cash flow as accurately as its assumptions, but is at the same time standardized, documented, and productized. Everybody wants to have his or her own Excel model — I suppose I come off like that too — but the underlying problem is that as soon as you want somebody else to understand and use that model, you need to do about 10 days work explaining for every one day work on the spreadsheet. With a good standardized model, it’s win-win, and when you break it into lots of detailed variations, everybody loses.
Whoops, suddenly my comment is too long and I’m only starting. I’m fascinated by this subject. I expect there will be a trackback here too, I’ll probably continue this on my blog. Thanks for the post! Tim.
June 3rd, 2007 at 2:43 pm
Hi Tim,
Thank you for a fantastic comment. Very interesting argument for standardizing financial models. My only concern is that unless it can be customized, a lot of businesses (especially technology and other emerging startups) have substantially different assumptions, drivers, and revenue lines that might not be fully expressed in a standardized model.
Thank you for reading and I look forward to reading your blog as well.
Best,
Eli
June 3rd, 2007 at 4:13 pm
Modeling Financial Projections for a Startup…
A good discussion of the need for financial projections as part of building a start-up company. They should build an analysis of the expected costs and anticipated revenue of their budding business.How detailed should those projections be, how should t…