Thursday, 14 August 2008

Is innovation over-rated?

Innovators are always the heroes of the story. They saw the opportunity when no one else could see it; they persisted stubbornly when everyone said they were a fool; they suffered hardship but eventually defied the odds to make it big, and so on and so forth.

And innovation is great. But we hear very little about the (undoubtedly many) poor sods who think they’re seeing something but it really is just their imagination, who persist stubbornly and foolhardily with something that ain’t never going to work, who continue suffering hardship till they vanish.

Do innovating firms perform better? Evidence from academic research on the topic is certainly equivocal: it is very hard to find solid evidence that firms that innovate (e.g. obtain more patents) perform better.

It so happened that I had a database of about 1300 firms active in some segment of the pharmaceutical industry, and all their innovations. Note, innovations do not all have to be real blockbusters but could just be new types of products or applications. These new types really can concern truly novel drugs new to the world but more often they concern a new dosage of some existing drug, a new intake form (e.g. pills versus injections), a new application of the same drug (i.e. a different disease for which a particular existing drug might also work), etc. Thus, not all of them are very radical innovations, but they are all new enough for them to have required clinical trials before their launch.

Using these data, I first tested, through some fancy statistical methodology, whether such innovations contributed to the growth of the firms over the subsequent years. The answer was a clear “no”; in fact, innovators grew more slowly.

Then I thought “perhaps they are innovating because they’re running out of growth”, so I applied a different (and even fancier) statistical technique to correct for that. Still, the answer was a resounding “no”: innovators subsequently really had more trouble growing and it was a direct consequence of the innovations.

So then I thought, “Perhaps I am looking at the wrong thing; and I should not be looking at growth but at firm survival”. Therefore, I changed my (already fancy) statistical methodology to test the impact of innovations on firm survival. But no, innovators died (i.e. went bankrupt) more often than non-innovators.

Then I thought “ah, it must be because innovation is risky; innovators may be the big failures of the industry but they are probably also the biggest success stories (I should really have thought of this earlier… better not tell anyone…)”. So I used an even fancier statistical methodology to model not only the average survival probability of the innovators (vis-à-vis the not-so-much-innovators) but also their “variance”. But no… innovators really did fail more often than non-innovators and with very little “risk”! In layman terms: they died pretty quickly and you could be sure of that. No risk-return trade-off here: do not innovate and you’ll get higher return for less risk…!

So then I gave up.

Might the answer simply be that innovation really is not such a very smart thing to do for a firm…? It seems, as an organisation, your chances of success are quite a bit better, and with little risk, if you simply stick to your guns, or at least not try come up with the new stuff yourself (but just wait patiently to imitate it).

But if that is the case, perhaps we shouldn’t tell anyone… Because innovation really is great and, as a society, we need it. But if everyone finds out that you, as an individual firm, are better off without it, nobody might do it anymore… So let’s keep this one under wraps, and between you and me, alright…?


Christian Bieck said...

This does sound like a sort of prisoners' dilemma. Are the innovators tops in any (disirable) category at all?

Anonymous said...

A patent does not naturally lead to the commercial success. Actually, most of the patents are useless from a commercial point of view. So it might be good to use other indicator for innovation other than the number of patents to do the statistical testing?

Freek said...

It is true, of course, that many - perhaps even most (dependent on the industry) - patents do not find their way into end products. Which is why, in the analysis reported in the posting, I measure "innovation" through the number and type of end products, not patents!

Wally Bock said...

A provocative post. Part of the problem may be that "innovation" can mean many different things. It can be the logical incrementalism practiced at Toyota, lots of small improvements added over time. It can be a step-wise improvement in process such as what happened years ago at Ford when they developed the first Taurus. It can be a patent. It can be one of those "creative destruction" thingies.

If we can't define innovation clearly and we can't measure it, no statistical methodology, however fancy, will help us. One thing we do know for sure, though. "Innovation" is one of the buzzwords du jour.

Freek said...

Wally: All very true and good points. However, I think that, relatively, I am not doing badly here with my data. I measure actual new products (so not R&D input, or patents, etc.) in the pharmaceutical market. Moreover, given the availability of detailed information on these products, I am able to split them up into six categories, ranging from drugs which are entirely new to the world (class 1) all the way down to straight imitations (class 6: new to the firm but produced by others before). All true innovations - i.e. classes 1-5 - produce negative effects for the firms. Only category 6 - straight imitations from other firms - can sometimes have a positive influence on the adopting firm.

Not perfect and not exhaustive, and limited to product innovations in the pharmaceutical industry, but relatively not too bad either I'd say!

thomas.j.hallam said...

I believe you may have missed the problem completely by using fancy 'statistical' tools.

In the Pharma industry, much similar to the Bioscience/ Electronics/ Internet and any associated industries 'Innovations' do occur by risk takers working in smaller and nimbler organisations.

However, survivability and success for small organisations is difficult. They cannot afford to fund/mass distribute their products. Due to this, after gaining scientific approval and associated patents the 'innovators' are often gobbled up inside larger firms.

This is a standard exit strategy for those scientists looking to make a return on investment. It also fuels survival of your larger and 'less innovative firms'.

Freek said...

Thomas: Yes, I am very aware of this. Fortunately, given the availability of detailed information on all these firms and all their (new) products, including which drug is transfered to whom, firm size, etc., in the statistical analysis it is not so difficult to correct for all these issues.

The way to interpret the results is, for instance, "if a firm of this particular size innovates, does it increase or decrease its chances of growth or survival?" In all cases, introducing innovations (no matter how radical or incremental) had a negative influence.

Wally Bock said...

My apologies. I did not mean to imply any weakness in your data, reasoning or methodology. In fact, I think the basic question you posed, "Is innovation over-rated?" was answered. My frustration is with "magic stone" terms which, if rubbed generate unending profits with little effort. "Innovation" is clearly one of those. "Leadership" is another.

What I liked about this blog and got me started reading it regularly is that you seem to ask questions that need asking, whether or not an answer is possible. This was a good one.

Wally Bock said...

Congratulations! This post was selected as one of the five best business blog posts of the week in my Three Star Leadership Midweek Review of the Business Blogs.

Wally Bock