Regular
posted 9 Jan 2008 in Volume 2 Issue 2
Opinion
Rolling forecasts
It’s 2am on a Tuesday night in March and you have just got the first complete run of your budget for year commencing 1 May. You feel tired, your wife thinks you are having an affair with your computer or worse, your assistant accountant. Your mind is between buzzing from the fifth coffee you have had this morning and the overwhelming desire to sleep. Oh, and the air conditioning went off at 9pm.
You nervously link the resultant profit and loss (P&L) account just delivered by your management accountant into your Excel spreadsheet to see what the basics of the budget look like for the first time: fees per fee earner, the value of a point; does it grow sufficiently to satisfy partners? Does it make sense? Am I really going to recover those fee increases and sell 15,000 more matter-related hours? Does the capacity increase match what’s happening with fees... It’s all too horrible and inside you are worried that some calculation is wrong and you have just produced absolute rubbish... You feel ill...
Well, I think we have all been in the situation described above at some time or another; I certainly went through something like this the first time I did the Watson, Farley budget some eight years ago (it was produced to the wire as office and practice heads did not fully commit to producing and delivering the figures). Truth is, 80 per cent of us still use spreadsheets, not a bad thing if they are clever ones. Over time I have acquired a suite of them with some visual basic, which can, in ten minutes, make me a budget or forecast with lots of ratios and key performance indicators (KPIs). As we all know, getting the information to feed the models is the difficult bit, but at least we all hopefully have okay systems to compile the final thing. So why go through all this pain more than once a year – even if it is much less painful now – and start producing rolling quarterly forecasts?
Back to basics for a moment, as we need to understand what the difference is between a budget and a forecast. Well, a budget is a specifically achievable target, which the organisation intends to achieve; it is usually fixed and for one financial year. A forecast is a prediction of the future in the medium and long term for the organisation, often based on KPIs and ratios going forward.
At Watson, Farley & Williams, we produce three versions of the year-ahead budget and create the final one in early July, just before we use it to measure our first two months’ trading. This is a very accurate budget, especially on the cost side, and contains all the last- minute changes and real happenings of months one and two. There are three main models underneath the complete budget:
1. A fee-generation model, fee-earner details, salaries, rates, capacity, qualification dates, effect of agreed client or matter discounts, etc;
2. Support salary headcount and salary increases;
3. All profit and loss lines with feeds from points one and two for fees and salaries. This is our main model and it accepts the same block of data that we use to make monthly accounts from our CMS Open system. It also outputs a very similar data format to make a new budget or forecast.
The three models have ‘tabs’ at our lowest granularity level – that is, practice group within office.
In addition to the budget-model suite, we have for the past three years, each summer, created a three-year plan: year one is budget and years two and three are predictions – that is, forecasts. It is quite aspirational and makes us think in the medium term about where we are going. One year drops off every year; it has been our ‘test bed’ for developing rolling forecasts.
We have a set of programmes and identical tabs for each practice group within the office. So, why not build a little bit more logic into them to move forward a quarter? The models take the data block that made months one-to-three accounts and exposes them to the forecast/budget model. Each P&L line can be amended by a budget holder or allowed to be predicted forward by just leaving an indicator set at 'predict forward' or some other criteria. The formula and structure of each granular piece (individual tab) – for example, litigation
-
Simply overwrite the four-12 and one-three of next year;
-
One-three actual divided one-three budget and then change four-12 and one-three next year;
-
Enter a percentage increase to uplift four-12 and one-three next year;
-
Base four-12 and one-three next year on other lines – that is, make NI be times percentage of all salary lines;
-
Other criteria.
The point is that with worksheets that really are identical, you are not supporting changes to say just the
After all this effort, why a rolling forecast? Well, at the end of the year you will find that you already have your next-year budget in detail, and that you have engaged budget holders in the process over the course of the year, not just once a year. Also, you will have been gifted a very accurate current estimate for the year every three months. You will also find that partners feel more certain about the future; they can see their success hopefully rolling forward.
I am presenting at the Managing Partner one-day conference on the 22 January 2008, 'Financial Planning, forecasting and budgeting'. If you attend, you will hear more about how to make rolling forecasts work and how to use what you have to make the process as easy as possible.
David Greenwood is director of operations at Watson, Farley & Williams. He can be contacted at dgreening@wfw.com.
denotes premium content | Jan 9 2009 







