Storing heat in a big water tank is certainly an option but it has drawbacks. I'm looking forward to developments in thermal storage in phase change materials and also storage via endothermic reactions.
They are insulated like a thermos (but more efficiently), and since they are really big, they have small surface area compared to their volume.
It is a VERY efficient way to store heat over shorter periods, like a week or so, and it helps making power plants more cost efficient.
The best way to make it efficient for long term storage (like seasonal storage) is to
1. Make them huge
2. Make them ball shaped
3. Put them underground
There has been talk of tryibg this out in Denmark. Almost all our powerplants (i.e. the not-really-old ones) are combined heat and power plants, and because we use district heating water to create a vacuum behind the power producing turbines turbines (instead of sea water) we go from a theoretical limit on the energy efficiency of around 60% to 98%, as we do not pump out the excess heat into the sea.
Newer combined heat and power plants have an energy efficiency of around 95% to 96%. The trade of is a slightly lower vacuum, meaning less production of power. However, we can scale up heat production to reduce costs of running power plants when energy prices are low (or even negative).
That more or less puts it out of reach for homeowners, condo dwellers, etc. You need a big expensive coordinated project to get that built.
Meanwhile, as an analogy, solar panels are within reach for a homeowner. Seasonal thermal storage will need to be more modular / modest sized for it to be widely applied. The potential for storing heat by forcing endothermic reactions (where you later extract it by running the reaction in reverse) has a lot more potential to be "homeowner sized".
Retrofits are possible, though with compromises to both extant structures and building envelope and passive energy systems.
Since efficiency of storage scales with size, community-based (neighbourhood-scale) thermal storage is an option. This allocates storage across a number of local structures, at the scale of tens to hundreds of structures per storage structure.
Similar notions apply to electrical storage, e.g., neighbourhood battery facilities. This works well for battery designs (e.g., liquid metal, molten salt), which are too technical and risky for safe household deployments, but could be deployed in clustered units with dedicated technical expertise.
The entire tank is packed within gobs of insulation. Cheap bulk insulation, rather than vacuum insulation, is effective, and additional volume accomplishes what the more expensive though thinner alternative delivers.
I worked for a major Danish power company for 2.5 years, but I mostly did mathematical modelling and implementation of the production planning system.
There are phase-change and transformational materials. Most complicate the process markedly, and may degrade (through loss or contamination) over time.
Only liquid-solid phase-chane is likely to be useful as liquid-gas volumetric expansion tends to 1:1000, leading to very large volume, or high pressure, or both, considerations, with corresponding costs and risks.
Aqueous thermal storage is remarkably inert. Small leaks are harmless, large leaks leave no long-term toxic legacy, pressures are ambient, materials are mundane, systems, monitoring, and controls simple, well-developed, and well-understood.