Our approach to adspend forecasting
Analysis of data over the past 30 years shows a clear relationship between the annual changes in total UK adspend (excluding recruitment) after adjusting for inflation and the annual changes in GDP. In fact, the changes in GDP account for around three quarters of the changes in adspend if the dotcom crash year of 2001 is excluded. At the quarterly level, the relationship is less clear as there are a number of short term factors that have a bearing on quarterly adspend levels. These include global sporting events such as the Olympic Games and World Cup and other influencers such as the timing of Easter. As a result, our adspend forecasts are based on annual GDP forecasts and for the latter we tend to use the consensus forecast published by the Treasury. This consensus is produced monthly and is based on around 40 independent forecasts from both city and non-city organisations. Other more subjective factors are also taken into consideration in our adspend forecast. For example, after the severe recent recession, some advertisers might well be cautious when considering increases in marketing budgets and wait until the recovery is clearly under way. This would result in a lower rate of adspend growth than would be otherwise expected in the shorter term followed by a higher rate in the longer term.
The forecast annual adspend total is broken down into quarterly figures after considering short term factors such as those identified above. The quarterly totals are then split into the display and classified sectors by examining the long-term share trends for each. Each element is then split by media type, for example television within the overall display sector, again by analysing the trend in the medium’s share of the sector. In all cases, a number of short-term factors will also be taken into consideration.
For recruitment advertising, which accounts for less than 3% of all adspend, we have a separate model which links changes in adspend to changes in the level of unemployment. A number of structural factors are also considered here such as the trend in favour of online ad expenditure in place of print.