Explaining significant changes to ICP market share in market segments, Feb – Apr 2020

  • 1.4K Views
  • Last post 14 August 2020
Matthew Keir posted this 13 July 2020

We’ve received several questions around changes in market share, market segments, and switching numbers specifically in relation to Meridian. I’ve published a simple dashboard that helps highlight what is going on. Care is required when interpreting the various market share measures, given the changes that have occurred.

Retailer growth in industrial ICPs, Feb - Apr 2020

Market share, as measured by ICP counts, can change due to:

  • winning or losing ICPs through customer switching,
  • winning or losing ICPs through commercial transfers or bulk switches (e.g. purchasing the retail book of another retailer),
  • winning new connections or having connections disestablished (e.g. ICPs changing status),
  • within a market segment, market share can change as ICPs are assigned new ANZSIC codes.

Meridian has shifted ICPs between market segments

Shifting ICPs between market segments occurs due to changes in ANZSIC codes for each ICP in the registry, and is not affected by switching. Early in 2020 Meridian migrated back-office systems. Their old system recorded an ANZSIC code for each ICP and allowed the business classification to reflect the activity occurring at each ICP.  Meridian’s new system records a single ANZSIC code per customer. In this situation, a single customer with an account pertaining to multiple ICPs will have the same ANZSIC code for all of that customer's ICPs. The shift observed in the data may be farm residences that were previously recorded as residential connections now being recorded as commercial or industrial, as the electricity account is settled by a business customer.

ANZSIC codes are required to be populated in the registry in Schedule 11.1 9(1)(k) of the Code. The provision refers to ‘customers’ at the ICP rather than electricity use “except as provided in subclause (1A), the relevant business classification code applicable to the customer at the ICP, in accordance with business classification codes published by the Authority”. Meridian did clarify this interpretation of ‘customer’ with the Authority at the planning stages of their migration and the Authority confirmed the interpretation. Earlier guidance from the Authority (at the time the Part 10 changes were implemented) sought to clarify the situation where a single ICP supplied multiple uses. In this case, the ANZSIC code populated in the registry should reflect the Trader's expectation of the predominant use of electricity.

The changes Meridian has made have a significant impact on all metrics presented by market segments on this website. During the time Meridian was undertaking their back-office migration, they were also active in winning customers through switching, and also winning new connections. This has made it difficult for some viewers to understand what is going on with the market share metrics. The public dashboard helps explain what has occurred. 

The public dashboard has four pages:

1. Market share trends, all ICPs and by market segments, 1 Jun 2018 – 31 May 2020

  • Meridian’s market share of all ICPs has increased over time to almost 15% (similar to the small and medium retailers combined).
  • The effect of Meridian changing the ANSZIC codes of ICPs en masse is evident during the period 1 Feb to 30 Apr 2020. The drop in residential ICPs is accompanied by an increase in commercial and industrial ICPs.

2. Market share summary, all ICPs and by market segments, 1 Feb – 30 Apr 2020

  • Meridian gained 6122 ICPs. If 5612 were from switching (see below), then ~510* ICPs must be the net result of new ICPs less disestablished ICPs. 
  • Distorted Meridian loss of 5.7% for their residential ICPs during the period despite winning 2112 through switching.
  • Distorted Meridian increase of 16.5% commercial ICPs and 30% industrial ICPs during the period.
  • Given the increase in ICPs in each business segment (see below) other retailers may have lost market share in the business segment despite not losing (or even gaining) many ICPs.

3. Switching summary, all ICPs and by market segments, 1 Feb – 30 Apr 2020

  • Meridian gained 16,368 and lost 10,756 ICPs giving a net increase of 5612 for the period.
  • Meridian had a net gain of 2112 residential ICPs through switching 
  • Meridian had a net gain of 2362 commercial ICPs through switching 
  • Meridian had a net gain of 1138 industrial ICPs through switching 

4. Growth in ICPs, all ICPS and by market segments, 1 Feb – 30 Apr 2020

  • Overall ICPs in the retail market increased by 6144, with 4720 ICPs in North Island and 1424 in the South Island
  • Distorted drop in residential ICPs across most regions – e.g. almost a 7% decline in residential connections in Ashburton
  • Distorted increase in commercial and industrial ICPs across most regions – e.g. 25% increase in industrial connections in Central Canterbury

This issue will also impact other metrics such as residential savings totals, residential consumption statistics, business demand, distributed generation statistics (e.g. residential ICPs with PV dropped for the first time ever in April 2020), and potentially other government reporting on the electricity sector that relies on these data.

* Switching numbers are recorded on the date the switch is initiated in the registry and not on the completion date when any switches impact market share. This difference means that there is a slight lag between the two metrics with switches initiated at the end of one month able to impact market share in the next month. Given this lag, the simple subtraction used to determine the contribution of ICPs to Meridian's market share during the period is only approximate.

msouness posted this 14 August 2020

ANZSICs are used for a number of non-retailer functions, including Low User tariff validation (may it eventually rest in peace) as well as network tariff design and validation of price categories.  

Is it worth putting in a Code change request to reword to "installation" rather than "customer"?