Win-back indicator summary

Report notes

1. On 31 March 2020, Part 11 of the Code was changed to ban retailers from contacting customers with an incentive to stay after those customers had initiated a switch away. However, customers may voluntarily choose to contact their previous supplier to withdraw a switch away, or to arrange a switch back. Such customer-initiated actions will appear in this report as potential win-backs, even though they are compliant with the Code amendment on saves and win-backs. The ban on win-backs applies to customers that were supplied by the retailer (or a retailer in their parent group) during a switch protected period lasting 180 days. The win-back ban is defined by the relationship between a retailer and a customer. The registry currently holds information on traders and ICPs, so looking at this information in isolation is insufficient to tell if an ICP-trader switch withdrawal or switch back is a retailer-customer win-back. Insights can nevertheless be derived where the report highlights outliers that may require further investigation.

2. This report presents metrics for win-backs via a withdrawal of a switch lodged by the losing trader, or by a switch back to the losing trader. Since 25 June 2015, a withdrawn switch at the customer's request must have a reason recorded against it such as ‘customer requested (CX)’ or ‘customer error (CE)’. Prior to the introduction of these withdrawal reason codes in 2015, a single code, ‘customer request (CR)’, was used.

3. The ‘ICPs won by’ parameter allows users to investigate the mechanism by which a trader-ICP win-back is executed, and the reason code given if it was via a withdrawal. For this report, ICPs won via a withdrawal process count the same as ICPs won via a switch. This approach is different from other market share reports which count completed switches.

4. The ‘Show’ parameter allows users to put the numbers of potential win-backs within the context of the scale of the retailer's activity or market share. The withdrawn switches lodged by the losing trader can be presented as a proportion of initiated switches away from that trader, showing how many ICPs may have been incentivised to stay. The mechanism for winning back an ICP can also be represented as a count or proportion of the total ICPs won by that trader (including via withdrawals) during the same month. This measure provides context for how significant the method is to the trader.

5. Switch types are included as a filter on the report. Customers are obviously free to move house and thereby become associated with a different ICP at any time. We are unable to assess win-backs of customers after such a move. Most interest will be in Trader switches, where the customer has remained at an ICP and has chosen to switch suppliers. Move In switches that appear to be won-back may be the previous trader at the ICP competing to retain supply of the ICP after a new customer has moved in. This behaviour is not banned but is included for completeness.

6. Market segment breakdowns are available from September 2013 when ANZSIC codes were introduced into the registry in a field maintained by the trader. No breakdown is available at NSP level. Market segments are not mutually exclusive. Residential connections have no ANZSIC code. Small and Medium Enterprise (SME) connections are defined as those assigned meter categories 1 and 2 (low voltage up to 500A) with an ANZSIC code excluding those relating to central or local government and other utility services. Industrial connections have ANZSIC codes from A through E while commercial connections use ANZSIC codes F through Z. More information on ANZSIC codes is available at www.stats.govt.nz.

7. Retail entities are parent company or trader. A grouped option is also provided separating the largest 5 retailers from small and medium retailers. Switches between entities within a parent company are excluded from this monitoring (i.e. a retailer can't win-back from itself). However, a potential win-back is recorded for a trader if an ICP switches from another retailer, and the ICP was previously supplied by any trader within the winning traders parent group within the 180 days. Bulk ICP transfers between traders that don't involve consumers are excluded. Some care is needed in interpreting the results from retailers of different scales. Those with significant market share may have some bias given the increased likelihood they have previously supplied any given ICP. This is especially apparent when looking across time scales longer than 180 days.

8. Region types are either network based or derived from standard statistical boundaries. Network-based regions are network supply points (NSPs), network reporting regions, and zones. Mappings are available in the NSP table which is available as a wholesale report. Other regions are derived from ICP location data and statistical boundaries (2013 census). ICPs without location data default to the location of their NSP. Main centres use Territorial Authority boundaries (City Councils) where Wellington includes Porirua City, Upper Hutt City, Lower Hutt City, and Wellington City. Many data reports allow for the selection ‘All regions’ (where the regions are not the data series) to get data for all regions in a single file. Regions are listed in a North to South ordering, except for NSPs which are listed alphabetically.