What factors make you change your online shopping habits? Perhaps the recent spate of bad weather may have influenced you to purchase more of your shopping online rather than brave the snow and ice. Alternatively maybe this gave you concerns over whether your shopping would be delivered? Is the current economic climate reducing your overall retail spending and is this the case for both high street shopping and online? Today (Friday 22 January 2010), Capgemini in collaboration with IMRG (Interactive Media in Retail Group) issued a press release predicting that e-retail sales value will continue to rise in 2010 but at a slightly lower rate than in 2009. The IMRG Capgemini E-Retail Sales Index is a tool for tracking the growth in UK online sales value, by monitoring a basket of retailers’ online sales activity. The index was founded in 2000 and today has over 100 retailers participating in it. The index shows that e-retail sales are up 14% in 2009 on 2008 and we are predicting that in 2010 this growth will be 13%. This edition of Figure It Out reviews how Operational Research has helped to contribute to this forecast; looking back at our analysis for 2009 and looking into the crystal ball for what 2010 may hold for e-retail sales. Our 2009 Forecasts In January 2009 we did some analysis to produce a forecast of e-retail sales for 2009, the graph below shows the accuracy of these forecasts compared with what the index actually demonstrated. IMRG_Graph 1.jpg On the whole our forecast was a slight over estimate but with a very credible overall average percentage error of just 4.5%. This shows that on the whole forecasting techniques can be a very good indicator of future performance. Some observations on where our forecast has deviated from the actual index value include:

  • October, where postal strikes led to reports of 10’s of millions of items undelivered or delivered late, resulting in a loss of confidence in the e-retail channel and a reduction in the index value in the final two weeks of the month, which typically see an increase in e-retail sales.
  • April, where e-retail sales fell on March due to the fact that Easter and Mothers Day fell one month earlier in 2009, resulting in e-retail sales falling in a month that typically has a small increase in e-retail spend.

Therefore when forecasting there are always external risks which can affect the accuracy of the analysis; such risks need to be identified as early as possible and incorporated into any forecast. 2010 Forecast Looking back over seasonal and annual trends in the index value enabled the generation of forecasts for 2010. The graph below shows the IMRG Capgemini E-Retail Sales Index values over the period April 2000 – December 2009. IMRG_Graph 2.jpg Two particular facets of the e-retail monthly sales data are:

  • Trend – an overall long term upward growth in the index.
  • Seasonality – there are peaks within the index at certain points in the year (most notably Christmas) and over time these peaks have become more prominent.

In this knowledge we are able to use analytical forecasting techniques to project forward the time series, this gives the following results: IMRG_Graph 3.jpg The red line on the graph shows the forecasted values that the model has fitted, against the actual index values and forward into 2010. The blue lines represent an interval in which we are confident that the actual index value will fall in 2010. Our model has forecast 15% growth in the IMRG Capgemini E-Retail Sales Index in 2010, which would equate to approximately £57 billion being spent online. What are the risks that could impact the quality of the forecast in 2010? Of course there are other social and economic factors will have a bearing on these forecasts. Following the general election in 2010 there is the wide spread expectation that levels of taxation including VAT will need to be increased to cut the national deficit. This coupled with a continued rise in unemployment rates, is likely to result in us having less disposable income. There is also the question of whether we are approaching maturity in the number of people shopping online. Figures collected alongside the index show that like-for-like numbers of unique visitors to retailers websites increased by 42% in December 2008 compared with December 2007, whereas this fell to 15% in December 2009 compared to December 2008. The above points suggest that it might be appropriate to revise the forecasts downwards. E-retailers have other levers to pull, however, to counter-act such factors: improving web design and innovation to convert visitors to their websites into sales, for example. Conclusion 2010 looks set to be another year of double digit growth in e-retail sales despite the economy being in recession, with the online channel likely to be an important player for consumers to search for the best bargains. The forecasting model that we created predicted 15% growth in 2010, however 2010 looks to continue to be an uncertain year politically and economically which will affect the growth of e-retail sales. We decided to adjust the forecast to incorporate these factors and therefore we predict 13% growth in e-retail sales in 2010, which equates to £56 billion spent online this year.