Biases, imprecision and a ‘Like’ for Castle-Invicta

Or "Data tastes better with a pinch of salt"

In this post, I explain some pitfalls with certain market visuals, including my own.

And as a natural byproduct of that, I say why I think the Castle-Invicta deal makes sense for both sides and their customers.

My recent ‘End of Year’ posts proved quite popular and generated some interesting debates over both coffee and keyboard. It has been great to reach a wider audience, so thanks to those that shared them and to those have engaged with the articles and with me directly. 

It helped me realise a curious behaviour of mine in producing these visuals: as they’ve progressed, I say less and less about the caveats (worrying that I’ll sound like a stuck record [or whatever modern equivalent is!]) even though that's no less important now as before.

So I thought I’d give some focus to those caveats. Apologies to those very familiar with the market ... maybe jump past the terminology. But I hope this will help inform your interpretations of the data. And to make one particular point, I take a look at Water Choice (Invicta’s trading name) and why the acquisition by Castle Water makes great sense.

Some Terminology

What’s a SPID?

A SPID (Supply Point Identifier) is one service, either water or sewerage. The majority of sites have one of each. Hence, two SPIDs per site is a decent ‘rule of thumb’ but not definitive. Good, but imprecise.

Some sites don’t need a given service because they have their own supply, such as a natural well. Others are complex and the wholesaler may have decided to create multiple SPIDs to achieve the most accurate data set up. Then you have multi-occupancy sites. and so on.

‘Associated’ Retailers

The water companies that manage the networks mostly gave rise to a wholesaler and an ‘associated’ retailer (a market term showing lineage, since legal separation was not enforced). Some networks decided to sell their portfolios before opening but even so, you can say with decent confidence who the retailer was for any given customer on the day before the market opened by looking at the wholesale area a SPID falls in. 

Precise? Very close but not exact. There was competition already for large consumers (over 5 Ml) so a small number had already switched. This has a negligible effect on SPID analysis but would be significant for volume, given their high usage status. My ‘Head to Head’ analysis also excludes the very small networks but again this would have a minuscule impact.

SPID v Site v Customer

Two SPIDs can be formally related in MOSL’s CMOS system, called ‘pairing’, though, as explained above, that may not be what a customer may recognise as their property... you may need a cluster of SPIDs for that.

Of course, many businesses have more than one site so a ‘customer’ is something else entirely. A company with sites spread around the country would have had multiple water companies serving them, going into the market receiving many bills to manage.

A customer with five two-SPID sites spread around the country signing up with a new entrant would cause 10 SPIDs to get transferred, which would show as coming from the various incumbent retailers to the new entrant.

WOCs and WASCs

Since the water and sewerage networks are separate (thankfully!), they don’t need to cover the same geography and you get overlaps between the networks managed by different companies. 

Some networks are run by Water-only Companies (WOCs), while others are Water and Sewerage Companies (WASCs). Thus, properties in a WOC area always have an overlap with a WASC’s patch, providing sewerage services. But other overlaps exist, which can complicate any broad assumptions made. 

One complex relationship is between Northumbrian and Anglian’s networks - and it underpins part of why the Wave JV made perfect sense. Each are WASCs but they have water-only ‘islands’. Anglian manages a water network in Hartlepool, where Northumbrian provides Sewerage. Conversely, Northumbrian owns the Essex and Suffolk water networks, where Anglian is often the sewerage provider. There are similar attractions with Water Plus, where the United Utilities and Severn Trent areas share an extended boundary area, even if that is not so explicitly a WOC/WASC relationship.

Bias #1: Not all SPIDs are equal

Of course not all sites - and hence SPIDs - have the same value to retailers. A newsagent does not need as much water as a coffee shop, let alone a power station. Volume can be a better gauge of actual switching value, though profit is a local and private concern, ultimately

This bias favours large multi-site switchers, given that micro-businesses are under-represented in the market, relative to their numbers

Bias #2: Non-switch gains

Let’s say you have a large national company with 500 sites (made up of a neat 1000 SPIDs) spread evenly around the country. If that company switches to a new entrant, they all switch. If that company switches to an incumbent, however, a bunch won’t need to switch: they will already be registered against the supplier. 

The bigger the current retailer’s portfolio, the fewer the sites they would actually need to switch. Thus, national gains by Water Plus (which would switch 678 SPIDs in our example), Castle (816) and Wave (842) will be most understated in the overall figures. 

Yes you can argue that is still reflective - there is no customer gain at those sites after all - but any analysis that focuses on switching needs to keep this in mind. The underlying value of the portfolio has increased by having renewed the contract on those customers.

This bias favours new entrants over incumbents, relative to the size of portfolio

Bias #3: Non-switch renegotiation

A similar point to the last one is that any decision to stay with an incumbent (not just those in a national company) are going to sit outside the switching figures.

This bias favours new entrants over incumbents when looking at switches

Bias #4: New retailer over consolidation

For most business, unless you have trade effluent, your sewerage retailer is largely a 'rent taker'. Your sewerage volume is often a function of water usage and provided all other sewerage charges are reflected accurately by wholesalers, there is little for the sewerage provider to do on your behalf on an operational basis. 

Given that, it often makes sense for companies to have a single retailer for both services: they can probably negotiate a discount and they have a single bill and relationship to manage.

Hence, if you do have different retailers, it makes sense to move to one, whether by consolidation to one of your current providers or to go elsewhere. If you do the latter, both services switch, whereas only one will switch if you are consolidating.

The data suggest that moving to a single retailer is a string driving force but consolidation would get under-represented by a ratio of 1:2 versus switching to a new retailer, all other things being equal. 

WOCs are the most heavily invested in this kind of transaction, which we’ll look at more closely below (see WOC jeopardy and Castle-Invicta). On one hand, their transfers tend towards 1 SPID for these transactions but on the other they are more frequently involved (as either loser or gainer) in switches than single-retailer sites.

This bias favours new entrants and former WASC incumbents over former WOC incumbents, although WOCs appear disproportionately more often in switch transactions

Bias #5: Zero sum games

Maybe when Bob Dylan sang ‘When you’ve got nothing, you’ve got nothing to lose’ he wasn’t thinking ahead to water competition but in a zero sum game, there have to be winners and losers. Yes, incumbents can (and do) win from each other but any gains by anyone are going to be at the expense of incumbents generally. 

A new entrant that does not win won’t grow, but nor will they shrink. There can’t be any negative drag on their momentum. Our own Retailer Grid does weight towards more recent transactions but new entrants will still tend to be positive, causing a negative drag on incumbents overall. 

That’s a fact of life in a competitive environment - the best antidote is to lose fewer and gain more.

This bias affects momentum visuals by placing more drag on incumbents than for new entrants

'WOC Jeopardy' and Castle-Invicta

I've discussed in other posts the inherent risk for WOCs in sharing all their customer relationships with at least one other retailer. Yes, that provides opportunity and of the two, the water retailer tends to have the better established connection. But WOCs are smaller companies by portfolio size and a number have already been acquired, formed joint ventures or sold portfolio.

In Bias #4 above, I touch on the bias aspect that moving from a two-supplier relationship to a single retailer gives greater weight when done to an entirely new party rather than consolidation to one of the incumbents, but let's look at that from the incumbent WOC perspective in light of Castle Water's acquisition of Invicta.

Invicta (trading as Water Choice), has 51,323 SPIDs (of which 49,230 are water, 2,093 sewerage). Most of its 2,224 gains are sewerage (i.e. the 2,093), plus the vast majority come from its counterpart WASCs (1,372 from Castle; 732 from Business Stream).

Most of its 3,760 losses are more broadly spread, though Business Stream gains the most. In short, it is proving more likely that it will lose water than it will gain sewerage, given that it is defending against all comers.

Overall, if you assume that all sewerage gains are to pair up with a water service (even those newly gained water sites), then it's dual-service sites are 4% of all its sites (notwithstanding the various imprecisions in that statement). In other words, 4% of sites have been de-risked from that jeopardy.

For contrast, SES Business has managed to get that ratio to around 41% (by the same assumptions), though that is why we made it our Retailer of the Year, admittedly. Water Choice was not performing badly but nor was its risk profile changing significantly.

Through the acquisition by Castle, a significant portion of that portfolio can be unified under a single retailer, which will be beneficial to those customers - and of course any that do want to move away still can.

Castle has recent experience of on-boarding through acquisition, which can help the experience of both employees and customers alike. That process may take time but there is no doubt that the portfolio is more secure after the acquitsition than before.

This article first appeared on our website (www. brodick.net) on 8th June, 2018.

About the author: David Tyler is a data and utilities consultant of over 20 years' experience, working with a range of the UK's leading energy and water companies. His data expertise extends from the strategic to the analytical, as comfortable in the abstract as he is with complexity and detail.