I think they are probably cutting off their nose to spite their face. Consider this hypothetical but perfectly possible scenario. Note that these people are entirely invented for purposes of illustration.
Suppose that I have learned from a gravestone that my ancestor Alexander Macdonald died in 1920, aged 60.
I go to SP and search for m*donald, alex*, death in 1920, aged 60, and I find a result. So I order the certificate. From it I learn the names of his parents, and I note that his death was registered by J Smith, sister, residing in a tenement in Glasgow.
Using the IGI I find his date and place of birth, and also that he had a sister Jean, born in 1870. I am now curious to know about Jean Macdonald or Smith.
Under the old system I would search for m*donald, j*, other surname smith, date range 1920 to 1980, birth year 1870 plus or minus 3 years. Under the old system, I would get a manageable number of results, from which I would find the right certificate, possibly using several batches of credits to do so.
If I try the same search in the new system, I cannot include the year of death, only her age, which I would enter as from 50. This produces 12,978 results. It seems that the search is returning "m*donald OR smith" instead of "m*donald AND smith". Even if I restrict it to Lanarkshire, it returns 3,092 results. At this point I give up, and I use no credits.
So whereas the old system would encourage me to look at some certificates, the new system discourages me altogether, and my expediture actually goes down. I am also a dissatisfied customer because I know that the information exists and I cannot get at it.
On the brighter side, I see from their technical stuff page that they are looking into bringing back the year of birth range for death searches, which will be a great help.