Niche success viewed from an engineering persective

You can’t be everything to everyone. Especially if you are a service provider.

It is common sense to think that the more you concentrate effort, the more successful you will be at a given task – otherwise there would not be colloquialisms like ‘spread yourself too thin’ or ‘focus on core competencies’. The downside, as a business to nichifying your offering (“NYO”) is loss of potential customers, alienation, or picking the wrong kind of niche. it also can potentially increase the risk of the demand moving out of your scope of offering, since it is narrow.

Is there a way of quantifying this level of focus or generality? is there a method of determining the statistical risk “what focus where”?

This may be a problem in a small market or in a market with small margins. In any market there is a certain demand over a spectrum of subcategories – an example is the digital camera market, which has everything from point-and-shoot to $10,000 SLRs, waterproofing, movie modes and any number of variables. Reduce that for arguments sake to a two dimensional spectrum: demand versus niche:

Digital Camera manufacturers face a landscape like this. although the variables would make this landscape more than a spectrum (more like a hyperspace), an example simplification could be:

There are three areas of obvious demand, based on, say sales figures. But what may be misleading is that demand spikes and troughs are inherently affected by availability. Take the trough between compact point-and-shoots and superzooms. There is indication of very little demand. However, there are no offerings in this area. That is where user testing, prototyping, polls, and marketing decisions of risk vs. pay-off all come in. This is where a gamble comes in. The true demand curve may actually be:

(pngs work much better in worpress!)

There is no visibility into this area without a product to occupy it and generate sales numbers, but there vould be a large latent demand for a new niche of “compact superzooms” and in fact, this was the case – the niche is now being aggressively filled by most major manufacturers.

However, those major manufactures have effectively unlimited capacity to explore new niches – if demand is there to be found, it will eventually be detected and “mined”.

What about a small organization that would have to remove resource from ‘known good’ areas of business demand to take a risk on another peak or even a trough? The advantage of another peak may be that it is higher, meaning high demand. A wide plateau is a broad area of demand for similar services. a deep trough may be a total lack of demand (the Homermobile) or an unexploited peak. Where a trough turns out to in fact be an undetected peak, there is a huge opportunity for the first business to offer services in this area.

Pilot projects. Skunk Works. [Cute Software Name] Betas. These are all manifestations of this exploration for new niches. But which are successful and who are successful, are there general rules that can be applied to any niche exploration?

The short answer is yes, and these are gleaned from engineering principles related to control systems. If you can tune a radio, you’ve already mastered the fundamentals.

Think of the niche spectrums above as the FM dial. Think of popular niches (spike of demand) as radio stations, and troughs as dead air.

Consumers, customers, and clients are like listeners. Some love their one favorite station and keep it locked. Others have worn out the Scan button on their car stereos in the hope of catching a sing, interview, or traffic report that fits their immediate needs. Some have one station for music, one for news, and one for traffic.

Now suppose you are the new radio station/niche service on the block and you have to pick a place on the spectrum. Here are some options for success, assuming you have a fixed budget:

  1. put in a whole bunch of low power power transmitters across a broad frequency range. This minimizes interference from other stations, and makes it more likely that someone scanning the dial will land at your station. Think of this as spreading your efforts over a wide spectrum. The disadvantage is that any listeners have relatively low power to listen in on, and so you reach only nearby people
  2. put all your power into a single transmitter right at the same frequency as a popular station. For some if not all listeners, you can overpower  the incumbent station and their favorite radio preset now belongs to you. of course, the incumbent could just ramp up their power, and you are now in a race over the same group of listeners and may alienate them all with the infighting.
  3. focus all your power into a very narrow but powerful station between two other popular stations. This minimizes interference, however new listeners will have to actively find you, and stand less chance of doing so since your frequency band is so narrow. However, there is no interference from other station, and you have lots of power, so your signal could potentially reach far more listeners, if only they would tune in.

This is the same radio station, with several options and a multitude of variations. They are all valid. Option 1 has the best chance of finding new listeners that are unaddressed by other stations. Option 2 has the lowest chance of failure due to lack of demand because it is attempting to exploit a known group of listeners, and option 3 has the highest potential to become dominant in a new niche (if it exists).

The engineering concepts here that can be applied are Signal-to-Noise, and Q-factor.

Signal-to-noise is simply how well your offering is heard. Option 3 has the biggest signal power, with no surrounding stations (noise). It also has a very high Q-Factor, which is the measure of how pointy the spike is. Option 1 has the lowest Signal-to-noise (SNR) and the lowest Q, since it’s point is more of a plateau. Option 2 has a high Q, but since it is overlapping another station (noise) it has a low SNR. The colloquialism “making a point” has a lot  in common with this – you have to be loud enough to be heard above everyone else in the room, and specific enough to be understood clearly.

As a business exploring new niches, you may want to use this analogy the way engineers exploit new spectrum. look for a quiet are of the spectrum surrounded by large spikes. the surrounding spikes tell you that there is likely latent demand between, and the quiet area allows you to have a high SNR without a lot of power. If there is lots of noise everywhere, you must get your signal above the noise, and with fixed signal power, you can only do this by increasing the Q-factor – i.e. sharpening the spike. The analogy is that if you want to be heard by customers, you have to narrow your niche. You will lose listeners who are roughly tuned into you, but gain more where you have focused your power, assuming they are there.  may wish to start with a lower Q-factor (i.e. spread out your signal) to cast your net wide, then once listeners have tuned in, you can narrow and specify and boost power. we’ll call this “sweep and lock” – it is what your car stereo does to find stations when on scan mode.

So the methodology can be summed up as:

  1. Explore the spectrum of products and categorize them along a niche spectrum
  2. choose a quiet area of the niche spectrum between areas of high demand
  3. position your offering in this quiet area
  4. make the offering general enough to fully exploit the niche and catch as many unaddressed customers as possible
  5. using feedback from these customers, determine where in this broad area there is a spike of demand
  6. narrow your efforts to this niche.

In the web world, where it seems there are infinite users and infinite products, step 4 becomes counterproductive. In this model, there is always demand for a product that is specific enough, it is just a matter of reaching the people that want it. This indicates that step 4 should be essentially scaled back so that more effort can be put into making a big impact within your niche.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s