It's a good idea but elements of it could be mistaken as being close to pyramid/ponzi, where the promoter is considered to suck value from the fresh blood. New users might resent being bound for 20 years? If the fee was not within the usual transaction fee, they might feel they were being scammed and just create a new wallet to bypass the overhead?
I would be tempted to limit it to something significant to the promoter and trivia to the new user - 20% of the fees for a year.. or better the cost should come out of what the transaction fee would be anyway; so the delegates lose out and pay the promoter; that lower payment would encourage promoters to find really valuable new accounts.
So, the cost of this should come from what delegates are willing to pay promoters - and not from the new users. 20 years loss in fees seems like a lot.. I guess the advantage is it binds those people in but the period could be shorter for that too. My instinct is to avoid complexity and long term binding commitments that you might regret later for reasons that might include that the balance of those payments are considered off target relative to their real value.
It's certainly a good idea to push marketing's focus towards real actual change in the userbase and avoid the payments through dilution. The only worry might be that real useful PR and marketing that cannot claim the benefits might miss out but if those involved are rewards in other ways, perhaps they won't mind?
So, that is: pay from the transaction fee and not any overhead on new users; avoid the pyramid sharing of those fees for several generations or tiers; and pay a short term burst of real reward, rather than a long term commitment.