Tuesday, 17 July 2007

Economics of Software-as-a-Service

Talking to a former exec from an IM supplier to SMEs I gleaned some useful insight today. His company had sold a fairly standard type of secure IM to SMEs. To start off with they'd sold the software, installed on the customer's servers, on a rental model, mimicking the "Software-as-a-Service" approach which is now so popular with startups. However, he explained that the company had gained much more sales traction with customer when, after some time, they reverted to a classic upfront software licence and downstream support payment model. Whilst I could understand one or two customers feeling this way I was initially surprised that it was such a transformational change. I was surprised that customers weren't fairly uniformly delighted with being able to link the timing of the cost to the timing of their benefit.
Without knowing the details I can only speculate, but it did strike me that the USP for this provider's product was really about helping them feel secure in an application, IM, where their particular customers were less comfortable (otherwise I'd argue that there are loads of similar alternatives). So to some degree they were selling to maybe the "late majority" rather than "early adopters". Maybe these people in particular could be said to be rather slower to adapt to new models for buying software, but also perhaps to be characterised by a different cost of capital.

Cost of Capital

In economic terms most SME's, but especially early-stage technology companies, have a VERY highcost of capital (I've seen it figured out to be around 80%), whereas the customer's cost of capital- particularly if it is itself stable/mature (i.e. late majority), maybe more like 12-20%. It therefore makes little sense from a purely economic standpoint for an early stage company to use a "rental" type payment model for software provided to a customer who's an established business. Perhaps more simply put, an early-stage company may have to pay precious early-stage venture equity rates for money to fund the working capital element of providing a "rental model".

Summary of differences

I've attempted to put together what I could think were the key selection differences between a "rental" model and an "upfront" model.

"up-front" model"rental" model
Favourable where vendor has higher cost-of-capitalFavourable where the vendor has lower cost-of-capital
Suits better laggard customersSuits better early-adopter customers
High investment can make for reluctance to change outContractual commitments can keep customers "locked-in"
"Capital" budget can be harder for vendor to accessContractual details on termination, rate increases etc can be harder to negotiate through purchasing
Upfront revenue particularly helpful where costs of acquiring customers are highRental revenues helpful where costs of supporting customers are high

(Please use the comments section below- I'm sure this is far from exhaustive- and I'll add your contributions into this table as best I can.)
You may be surprised to realise how strong the Cost-of-capital effect is in practice. Take the following example:

  • Company A sells their system on a rental model for £10,000
  • Company B sells the same thing on an upfront plus support basis for £15,000 upfront and £2,500 annual support charge.
The Net Present Value of both of these revenue streams for a early-start company with a 50% cost of capital is actually about the same (my workings are here), whereas I suspect that most of their customers would very much prefer to pay that little bit more for a perpetual licence!
As always your comments are very welcome.