It doesn't seem, on the face of it, like the smartest strategic move. You're a drug company, many of your most profitable treatments are coming off patent, the mooted replacements haven't produced, so you sack a load of R&D staff. Hmm.
The job cuts that Pascal Soriot has imposed at AstraZeneca seem so clichéd. A new chief executive arrives and tries to impress the City with what a tough guy he is by cutting costs, which these days means jobs.
Not everyone was impressed, although the overall reaction to Mr Soriot's strategy was favourable, at least in terms of the share price, which finished ahead.
Here's the method in the madness. What Astra is doing is sharpening its focus, concentrating on what Mr Soriot feels are the best bets.
From here on, Astra will concentrate on respiratory illness, inflammation and autoimmune diseases, cardiovascular and metabolic illnesses, and cancer.
There will also be a secondary focus on infection and vaccines and neuroscience, although Mr Soriot says investments in these areas will be more "opportunity-driven".
Thing is, while Astra's is streamlining, other drug giants have gone the other route, diversifying in an attempt to reduce some of the pressure on their research engines.
Still, let's look at one of the opportunities that Astra is hoping will be a winner. Moderna Therapeutics, a small biotech firm based in Cambridge, Massachusetts, uses synthetic messenger RNA to tell a patient's cells to produce various desired proteins. It's a potential alternative to gene therapy and could, in theory, persuade a cell to produce any therapeutic protein you could desire, which could be useful in treating a huge range of conditions.
Astra gets first dibs (for a five-year period) on drugs using Moderna's technology to treat cardiovascular and metabolic diseases, plus some exclusive rights for cancer treatments. It is paying $240m (£158m) up front, and Moderna could make a further $180m down the line.
This is one of the largest partnership deals ever seen between a drug giant and a biotech that has yet to bring a treatment to clinical trials. Astra is betting big on it coming off.
Astra itself represents something of a bet for investors.
From a financial standpoint, the three sets of job cuts that have been announced recently – and more than 11,000 posts have gone in just over a year – will cost the company $2.3bn in one-off restructuring charges but save an annual $800m thereafter.
Well, that should provide further fuel for spending on deals, and Astra might need them if it continues to struggle with its own drugs.
The shares are not all that expensive, trading at 8.8 times this year's forecast earnings. That rises to 9.1 times next year's forecasts.
Plans for a share buyback have been cancelled to give more flexibility for doing deals, but the dividend is nearly twice covered by earnings, and analysts think it is stable. The forecast yield is a more than respectable 5.8 per cent.
In October I advised holding Astra at 2,896.5p. You'll be happy if you did. It is probably just about worth staying in there, for that yield if nothing else. If Mr Soriot does get it right the pay-off could be very big. But keep a close watch on developments.
If you want drugs, I'd stick to GlaxoSmithKline, which is far less reliant on blockbuster treatments (thanks to that diversification). I said keep buying at 1,406.5p in October, and made the shares among our 10 to follow after they had eased back to 1,335p by 31 December (I'll update on the 10 over Easter). The shares have since jumped, and rightly so.
Even so, at just 13 times this year's forecast earnings, falling to 12 times next year, with a 5.2 per cent yield rising to 5.5 per cent, it is a more attractive and less risky buy than Astra. I'd still rate the company as a buy even now.
Another share on its downers but worth buying, from a completely different sector, is Bloomsbury, the publisher of Harry Potter fame. The boy wizard might be enjoying his retirement, with the film franchise also finished, but the company is more than coping.
Today it said it will meet forecasts in both digital and print books for the year, and at only eight times next year's earnings (for the year ending 28 February 2014) while offering a forecast yield of over 5 per cent, buy to put a bit of magic in your portfolio.