“£307,756 to buy average house” was the headline in my local free newspaper today. Interesting I thought, how is this number made up I wonder? Is it the mean, median or mode? Was it calculated using the information about recent selling prices or the houses currently on the market? If so then what determines an ‘average’ house – number of bedrooms, bathrooms etc? Unfortunately none of this information was provided so I found it difficult to form any real opinion on the sensational headline.
Particularly in the case of house selling prices averages can be dangerous. A ‘long tail’ of multi-million pound house sales can easily skew an average calculation based on the mean. Maybe these houses were excluded from this measurement or maybe the median (which ranks all the house prices in order and takes the middle value) was used. We don’t know because the article didn’t say.
Without understanding how a measurement is calculated it is difficult to base any sensible opinion or make any decisions on.
To avoid pitfalls such as this, here are four Straight Line tips to make metrics actionable:
1. Use the right measure for the purpose. For example, in this case the median is probably the best measure to use as there is likely to be a long ‘tail’ of extremely expensive houses. Usually this is also true when measuring the time to complete an activity.
2. State the context of the measure. What period of history was used, how many data points were included, why were any outliers excluded? Were any other factors impacting the results? What is the statistical significance of the result, is it within an acceptable range and can any variation be explained?
3. Have a relevant benchmark to compare to such as prior week or prior year. Is the measure improving or deteriorating against the benchmark?
4. Consider what other factors may be impacting this metric. These factors could include segmentation, price, seasonality, geography, special events or external factors such as political environment, market conditions, competitor actions etc.
When you follow these guidelines you have a metric which is clear in its meaning and wider context.
Now it is safe to determine what actions to take.