Market Statistics
- Introduction
All businesses try to measure their markets.
They have two main objectives
- First is to measure their own performance against that of their competitors in the market.
- Second is to form a basis on which to forecast the future size of the market.
The farm machinery industry has some particular problems in measuring and forecasting its markets these are not normally met in other areas of marketing.
- Why is the farm machinery market different from other markets?
The biggest problem is that machinery sales are largely dependent on crop production and this depends mainly on the influence of the seasons. Sales in a good season can be two to three times more than in a poor season. This is a much greater variation than is met in most markets. When sales fall, it is impossible to tell whether a product is losing market share or whether the market is falling unless there is effective measurement.
Few farm machines sell evenly all the year round. Many machines are only used at specific times of the year and can only be retailed close to those times. Combine harvesters, balers and hay tools are obvious examples but tillage and seeding machines and the big tractors that operate them are almost as seasonal.
Since manufacturing and shipping lead times are long, this puts a very high value on accurate forecasting and companies suffer severe penalties for making mistakes. If too few machines are brought in, market share is lost. If too many machines are brought in, heavy inventory holding costs are incurred, as the surplus machines often cannot be sold till the following season.
International commodity markets for grain meat, sugar, wool etc. also strongly influence farm machinery sales. Prices vary widely. Farmers do not set their own prices and a high proportion of farm production enters world trade. This is true of all farm machinery markets but is particularly important in Australia with its big area and small population.
- Why is the Australian farm machinery market different from other machinery markets?
The Australian farm machinery market has its own special problems because it suffers from more climate uncertainty than other big agricultural producers. On average, in about one year in five, drought reduces winter grain yields significantly. This is far more often than in the Northern Hemisphere farming areas where a crop failure one year in fifteen or one year in thirty is more common. These big falls in yield feed through into falls in income and hence in the demand for machinery.
The changes in weather patterns and rainfall over the last ten years have accentuated these problems and increased uncertainty.
In addition, the Australian market is not uniform. There are wide regional differences within Australia. Australia-wide droughts are rare, though not unknown. There is usually some area in Australia that is having a reasonably good season. It is important to be able to measure the effects of these differences.
Western Australia has lower average rainfall than the eastern states but the rainfall tends to be more reliable. Queensland and northern New South Wales have a high proportion of winter crop failures, but benefit from a rainfall pattern that allows both summer and winter cropping. Failure of both crops in the same year is rare, though again not unknown.
Australia has a very big climate range, so there are many different types of agriculture, ranging from low density open range grazing of sheep and cattle to intensive irrigated horticulture and cropping.
Because of these regional differences, national overall market trends are of little use in trying to measure and forecast the machinery market. It is necessary to look at each region and each activity separately to come up with a national picture.
Given Australia’s great size and relative sparseness of population, effective distribution of products to farmers is vitally important. Although other systems have been tried, independent dealerships have proved over time to be the most effective means of distribution. Good dealers are the key to successful farm machinery marketing and dealer areas are therefore the key to market analysis.
- What has the Australian industry done to deal with this?
In the early post-war years, the Australian Census Bureau ran a service that reported quarterly on retail tractor sales by state and statistical division in broad power classes.
Statistical divisions gave a reasonable good regional picture but it was still far too coarse a system to define dealer areas accurately. The power classes were very simple – under 10 Hp, 10-30 Hp, 30-40 Hp, 40-55 Hp and over 55 Hp.
Another Census Bureau service reported quarterly on the wholesale deliveries by state for combine harvesters and balers and other major items of equipment. There was no regional analysis of this information.
These systems provided a comprehensive coverage of the market for the time. However both reporting systems were slow and results were sometimes not published until many months after the end of the quarter.
Companies in the industry began to exchange copies of their census bureau returns during the 1960s, mainly so that they could get a quicker reading of what was happening in the market.
Computer analysis began to be used in the late 1960s and this made much finer analysis possible. A private system, independent of the Census Bureau, was set up in the early 1970s that recorded sales at end user postcode level this finally made it possible to measure and develop accurate market estimates at small area level.
At the same time the explosion in tractor power and types made more classes necessary this was done by recording the make, model and the power being sold. It then became possible to measure any power class, no matter how it was defined.
Combine harvesters and balers were included in the system on the same basis as tractors.
With sales being recorded by model, it also became possible to set up a full database of specifications. This was used to produce product specification comparisons databases.
As a general rule, it is best to use the finest specification and area divisions available in processing data. It is simple to aggregate these into broader divisions but it is impossible to break down the information into smaller divisions if that information has not been recorded to start with.
That is the origin of the current Agriview system that operates today.
2. Analysing the market
Agriview programs use four basic time periods to measure the market, months, quarter, year-to-date and rolling 12 months.
2.1 Time Periods
Monthly sales figures are useful to provide a snapshot of the latest state of the market but this has a number of shortcomings. They take no account of seasonality and this should be kept in mind when evaluating them. Some months are much more significant than others. January and February are notoriously unreliable as indicators.
Monthly Figures for states or dealer areas can be very small and may not be statistically significant. Monthly figures can also be distorted by individual company sales promotion programs, supply impacts due to factory lead times, shipping constraints and Government investment allowances just to name a few.
Quarterly figures provide a better picture of what is happening at a dealer level particularly when compared year on year but they still suffer from market distortions be it either supplier induced or seasonally induced.
Year-to-date figures give a more stable picture than monthly or quarterly figures, particularly as the year advances. They are particularly useful when working with a company’s own sales budget figures, helping management to evaluate why actual sales are varying from budget. Year-to-date results can be adjusted to match each company’s financial year.
Rolling twelve month figures usually give the best idea of a company’s performance in the market. They eliminate any seasonal factors and promotional and supply line effects. They give a good idea of the trends in the market, and again can be set up to match each company’s financial year.
Long-term analyses is very important in this market, there tend to be cycles in agricultural product demand – for instance, those caused by changes in the balance between livestock and cropping. These affect the demand for machinery. There are also long term trends within the machinery market – for instance power trends – that run over far longer periods than twelve months.
Five year graphs are often useful in understanding the market but charting more than ten years can be unhelpful because of the number of changes in circumstances that occur over such a long period tend to make the earlier figures irrelevant.
All numerical series can be easily converted into value series by using a machines retail price to generate new figures. This is worth doing regularly in a market where there is such a big gap between the cheapest and most expensive machine. Bare sales figures can be misleading.
The same can be done with power ratings to generate figures for total power delivered. This helps to put earlier market figures into perspective. The big increase in average tractor power coupled with the complete changeover to self-propelled harvesting equipment throws a quite different light on some of the apparently high earlier tractor sales figures.
2.2 Power Groups and size classes
Tractors are usually grouped into power ranges. This can be a trap when an individual tractor’s power is increased slightly, so that it just moves into a higher power range. The market can appear to change dramatically when it has not really altered at all. In practice power ranges need to be continually adjusted to allow for changes in the market.
The best way for a company to analyse the market realistically is to build the power ranges around its own product line and its knowledge of which products really compete with each other.
This can only be done if complete detailed information by model is held in the information system.
2.3 Geographic Areas
State figures are the first breakdown of the national sales figures. Traditionally states also tended to be the natural divisions between company branches but this has changed over the years as company branch structures have dissolved. While they are better than nothing, state boundaries do not necessarily follow natural trading boundaries and frequently split dealer areas, particularly along the New South Wales – Queensland and New South Wales – Victoria borders.
Regions can be usefully defined, based on different farming types, for instance the Darling Downs (for grain) and North Queensland (for sugar) in Queensland and the Murray and Murrumbidgee irrigation areas in New South Wales and Victoria. This can provide useful insights into the sales performance of particular models with appeal to specific industries.
Dealer areas should be the basic areas used for market analysis. Dealer areas are built up as the sum of a group of postcodes. While boundaries are not precise, quite accurate estimates of the actual dealer area can be made and have tendered to become more accurate as dealer areas have become bigger. Dealer territories can then be aggregated to sales territories.
There is a continuing tendency for dealer areas to increase and for the same dealer to handle a number of dealer areas. In addition, dealers are tending to take increasing responsibility for forward ordering of inventory. This is producing an increasing need to analyse market areas within dealerships. Manufacturers who encourage better dealer knowledge of market trends are likely to benefit from improved dealer performance, particularly in the critical area of inventory management.
3. Figures and Forecasting
The main application of market analysis in business is to produce forecasts. There is no choice in the agricultural market but to forecast. Not to do so is equivalent to forecasting no change in market conditions – the one forecast that is guaranteed to be wrong.
3.1 Building up a Forecast
We believe the best way of building a market forecast is from the bottom up rather than from the top down.
The analysis needs to be based on an individual assessment of each dealer area.
A five-year set of figures roughly defines the range of likely sales.
Evaluation of the seasonal outlook must be based on direct contact in the dealer area to assess order books, local confidence, weather patterns and competitive pressures.
In looking at crop prospects, water availability is the most important factor. The amount of water in major storages is a known factor and helps forecast the results of irrigated cropping.
With dry land cropping, it is more complex. The expectation must be that rainfall will be average through the growing season. The growing season is roughly April through November for winter grain crops and October through April for summer grown crops.
The timing of crops in Australia means that initial commitments for combine harvesters have to be made before there is any hard information available on the moisture that will grow the crop.
Continual updates are necessary. The Bureau of Meteorology website measures rainfall on this basis and maps precisely where rain has fallen on a daily basis. The likely effect of above or below average rainfall can be fed into the forecast as the season progresses.
The total sales forecast is the sum of all the individual dealer area forecasts. Given the uncertainties of weather and economics, it should be reviewed at best on a monthly basis or at worst on a quarterly basis.
3.2 Relating to the ABS Census System
Farm activity is measured in the agricultural census. The census used to be carried out annually but for reasons of economy, this was changed in 1996 to a system of five yearly complete census collections with sample surveys for the years in between.
Sample survey figures must be treated with caution, unless they come from very large areas and apply to major farm activities, when you get down into small areas and minor crops such as pulses or oilseeds, quite big sample errors occur.
Any fine analysis has to be based on the complete survey years, 1997, 2001, 2006, 2011 and 2016 with the sample survey years used to check major trends.
The current Census system introduces a further problem in that it is a matter of luck whether the census falls in a good or a bad year. This needs to be assessed before the meaning of the figures can be evaluated.
The sample surveys can be useful in helping to asses this.
3.3 Relating census data to dealer and sales territory areas
Census data is published based on the Australian Statistical Geography Standard (ASGS) 2011. There are five levels below State and Territory level, the SA2 statistical area level is preferred as it breaks the country up into 2,214 regions which for the most part allows better matching to post code areas which are usually smaller.
Approximate fits of dealer areas and SA2 regions can usually be developed quite simply.
It should be remembered that some postcodes originally signified mail delivery routes and cover quite large areas of country. This is particularly the case in Queensland with postcodes 4352 based on Toowoomba and 4702 on Rockhampton, covering very large areas of territory.
A good computer mapping program is essential for identifying and mapping dealer areas.
Local input about where a dealership actually trades is also essential.
Using this basic approach it is possible to set up an information system which identifies how many farms there are in a dealer area, what they do, how much cash income they generate and what machinery they buy.
3.4 Measuring the External Factors
It is much more difficult to measure and judge the external factors affecting a dealer area’s prospects.
Farms are part of the general economy and affected by booms and recessions just as much as other independent businesses. Events such as credit squeezes have effects on the farm machinery market just as severe as droughts. It is vital that any market forecast takes this into account.
ABARE (the Australian Bureau of Agricultural and Resource Economics) is the major source of agricultural economic information. However, it operates essentially on a national scale and is not necessarily much help in the sort of detailed forecasting needed in the farm machinery industry.
The exchange rate is a particular hazard both for farmers and machinery importers. A high Australian dollar threatens farmers because it reduces the value of their exports but favours importers because it reduces the import prices.
Weather forecasting is gradually getting better as weather patterns become better understood. However, it is important to remember that the Bureau of Meteorology only deals in probabilities. No one can yet forecast with certainty that there will or will not be a drought. They can only indicate that one is more or less likely.
Finally, politics must be taken into account. Agriculture is a highly political industry with powerful lobby groups pushing particular interest in most countries. Not all policy decisions are made on an economically national basis. This is true both in Australia and in world agricultural markets. Some lobbies, such as the USA sugar industry lobby have quite disproportionate power to distort world markets.
Assessing this obviously depends on the forecaster’s judgement. Even the best set of market figures cannot help.
GOOD LUCK!
