Estimating a negative binomial curve, from sample size and mean

Hey guys!

Apologies for the potentially *stupid* question... I'm *very* rusty on my maths and am not getting anywhere with Google's support.

I'm a marketeer, looking at how/if it may be possible to estimate a NBD curve from a limited dataset... my client has given me some 'data' about customer use:

Total number of transactions (34,300,000)
Average number of transaction, per customer (12.4)

They have not given me (nor do I think they have gathered) any sense of any additional variables.

Looking through the various calculators, papers, threads and other sources, I'm struggling to get a lay sense of whether I can do a rough regression model.

I could guestimate the max number of transactions per customer and I know the minimum is 1, in order for them to be seen as a customer.

Other than that, I'm struggling to work out how I may be able to model a regression curve that shows how far removed from the 'average' a 'normal' customer is.

Any support would be most gratefully received.