Bollard's more conservative estimate is 38 hen-years per dollar, if you include other expenditures on farm animal welfare. I think we need to do include those because we didn't know in advance which efforts would be effective, and probably there will be some regression to the mean.
If you're thinking of this as an offset (instead of just directly comparing it to other charitable expenditures) then you need to credit other inputs - especially, the time of the people working at these places. Labor captures about 60% of national income. People working at a charity are often taking a substantial pay cut relative to what they could get on the market, so let's say they capture 30% of the value. That means that for every \$1 paid to employees of a leanly run nonprofit, the measured economic opportunity cost is a bit over \$3 worth of GDP.
Then there are the volunteers and people mobilized by CiWF. If we assume that each staff hour corresponds to about a similarly valuable commitment of volunteer hours (I genuinely have no idea if this is way too high or way too low) then each \$1 spent on salary has an opportunity cost of about \$6.
About half of CIWF-US's expenses are salaries. Let's assume that this is true for their overall organization. Then for each dollar, \$0.50 goes to salaries, with an economic opportunity cost of \$3. The other \$0.50 goes to other expenses, which we'll assume are rents, supplies, etc., all purchased at market rate from for-profit businesses capturing a normal return on capital and labor, so let's say the true opportunity cost is \$0.50. Thus, donating \$1 to CiWF actually costs \$3.50 in economic value, so you should only get credited for about 38 / 3.5 = 11 hen-years per dollar.
That's still a pretty high number. But how does the suffering alleviated here compare with the suffering inflicted by a year of eating factory farmed animal products? Let's say that about a third of the net suffering laying hens do comes from the cages themselves.
It seems like the vast majority of farmed animal suffering measured in animal-days (not adjusting for brain size) comes from poultry and eggs, since chickens are so small - and most slaughters will be attributable to such animals too, for similar reasons. But maybe broiler chickens suffer more than ordinary chickens. I can imagine being indifferent to spending ten days as a caged laying hen and a day as a broiler chicken (though I suspect the true ratio is smaller). So, on that model, a dollar alleviates suffering equivalent to about a third of a year of a broiler chicken's life.
If you're eating about a kg per day of meat, and about half of it is chicken, you're responsible for about 13 days' worth of broiler chicken suffering per day (assuming away elasticity considerations in the long run), i.e. 13 years' worth of broiler chicken suffering per year, so you'd need to give 13 * \$3 = \$39 per year to offset that diet. If you're not eating much chicken, it would take much less.