The analysis described in this paper is different from our other examples. It uses panel data, which is under-represented in this both. It treats advertising in the simplest possible terms - as cash expenditure. There is therefore little opportunity to make deductions about its quality, though we see that some combinations of brand and campaign (a useful reminder here about the product) give one brand an advantage over another. The paper draws overall conclusions from a large number of product fields, rather than looking in depth at a single problem. Its conclusion is an important one: that growth in sales share is associated with advertising expenditure. We cannot write is caused by because it is possible that the higher expenditure was sometimes planned at times when sales increases were expected for other reasons. Some people may see this comment as over-cautious; the reader has to make up his own mind on this point, preferably helped by similar analyses about a market he knows. The method of analysis was employed earlier by Unilever under the name dynamic difference.