Just because it's a client requirement doesn't mean you are allowed to ignore USPAP and potentially create a misleading report.
I recently ran a six months of sales (84 transactions) for an area of very conforming tract homes. The SP:LP ranged from 92% to 121% with a median of 103%. To further break it down: 26 sold under 100% SP:LP; 50 sold from 100% to 109%; and 8 sold for 110% to 121%. 35 of the 84 were 5% higher or lower than the median. Additionally, 77 of the 84 sales sold in under 100 days and the median was 38 days.
While a median or average can be calculated does not mean it can effectively be applied. Do you remember the stories of the two hunters who shot at a charging bear? One shot right by 5' and the other, left by 5' but on average they killed the bear. If I were in a casino, applying a median may be the best figure to wager a bet. But that's all it is, a bet with a margin of error as high as 18% in my case above. Of cource you could cherry-pick a listing to what you think would be the best property to apply your SP:LP adjustment however at that point the appraisal process is pointless.
You need to ask yourself at what point is an adjustment reasonable and credible? In the case I presented above the maximum historical error could be 18%. Breaking my data down further, 42 percent of the time the actual SP:LP differs from the median by 5%. When was the last time you applied a single adjustment which more than four out of ten times will result in an adjusted price of the comparable which had an error of 5% or more? At best, you could come up with an adjusted price and state the error potential based on historical data. Or, you could use a range for your adjusted price of the comparable with limits of the credibility based on histoirical data.
Personally, I provide listings but do not use a SP:LP adjustment. Rather, I explain that as of the effective date this is how the subject would rate against its immediate, non-cherry-picked, competition.