Regular visitors to our website may have noticed a couple of new charts residing there. We created them to provide some perspective on what the recent increases in our year-over-year metrics really mean to the “raw” or “absolute” on-line consumer demand for discretionary durable goods in the United States. Year-over-year comparisons suffer from two compounding issues that slip past intuition-based observers:

– Long term expansions or contractions generate compounded growth or shrinkage, which is non-linear and greater than intuition alone might grasp.

– Simple cycles in year-over-year movements that seem to offset each other (e.g., -10% followed by +10% and then followed by another -10% and then yet another +10%) are not zero-sum in absolute values — in the cited example the quantity being measured only returns to 98.01% of its starting point after the four sequential 10% down and up legs.

Our metrics are necessarily year-over-year for a number of technical reason, not the least of which is the need to remove an underlying bias in on-line shopping as the result of long term market-share movement away from brick-and-mortar retail. That “normalization” process uses “same shopper” metrics that rely heavily on behavioral statistics for IP address baselines that (as a practical matter) reliably extend back only 12 months or so. Additionally, year-over-year metrics inherently wash out seasonality issues — thereby preventing our data from getting hijacked (or creatively manipulated) by “seasonal adjustment” factors. For those and a number of other technical sampling reasons our data is (and must remain) year-over-year in its primary form.

That said, we still recognize that providing only year-over-year “rate-of-change” charts can mask longer term effects on the “absolute” levels of demand. This has been particularly evident over the past few weeks when our Weighted Composite Index broke back into growth territory — causing some media outlets to cite our data as further proof that the “recovery” was strengthening. While it is true that we have witnessed a recent surge in our year-over-year numbers (see below for more analysis), a reasonable perspective on the depth of the hole that had been dug by multi-quarter (indeed multi-year) contractions was completely missing from the headlines that our data generated.

Hence, we have created two new charts that follow our new “Absolute Demand Index.” The new index approximates the “absolute” values of on-line consumer demand, where a value of 100 represents the year-long average consumer demand for the baseline year of 2005. The first chart captures the new index on a daily basis for the past 60 days:

Related posts

• No related posts.