(note: this is an adaptation of my Marketing degree thesis and as such you may have seen a similar post somewhere else since I used it in a couple boards)

When dealing with marketing research and niches, we’ll always find ourselves in a point where applied research requires optimization to make the most out of our marketing budget(s). This point can be planned in advance (ie if we’re setting goals by time frames) or may surge out of necessity (ie, we notice our marketing efforts aren’t working as expected).

Now, if you’re worried this post may be something heavy and tiresome, don’t worry,  I won’t go into deep technical stuff nor research methodology. Instead, I’ll try to explain this in an as simple as possible way, by providing easy to understand examples and then some ideas on how to implement these techniques in a really easy way.

To start, let’s go with what I called the The Sniper and the Machine Gun models. In this scenario, we have two soldiers trying to reach their targets. One of them has a high precision sniper rifle, and the other has a machine gun. At this point I’m pretty sure you’re figuring out what will be this about, but let’s have a look at it:

The Sniper Model

This is the Sniper Model. Our sniper is using few bullets (money/resources) to make a sale by researching the proper way to get an easy target. Thus, the full line goes straight to the heart of the target (or, in other words, make a sale). Now, you’ll see the blue dotted lines that goes to the “body” of the target, “injuring” him, which may lead the target to “survive” (meaning he won’t buy because of a weak call to action or unrelated motives) or maybe he will “die” because of his injuries (meaning that he may NOT buy at that exact time, but may be buying sooner than later).

To add to that, our shooting range has 3 zones: niche targeting, media delivery (proper or wrong choice of media and marketing materials) and competition, and it’s easy to see our bullet will be farther from the target as the zones become bigger. The more accurate and restricted our “zones” are, the less loss in accuracy.

Let’s get the metaphor out of the picture and straight to business:  If we define our niche in an accurate way, our target will be closer to become an effective sale since we’ll have less dispersion. Once we start adding more factors, the dispersion ratio will have a compound rate increase.

Before continuing, let me add something that I know most of you hate to guts, but it’s very easy to use and extremely useful and needed in marketing and economics research: the Compound Interest Function. It’s as simple as this:
Value   =   IV (1  +  R / t)^Yt

where IV=Initial Value , R=Interest Rate , t= times of occurrence per year (every month would be 12) and Y=number of years.

In the graphic above, I’ve used a reasonable deviation of 3.4%, which comes out from the difference between total sales per campaign minus impulse buying multiplied by dispersion factors. So in this simple graphic, with only 3 dispersion factors, we could notice that around 30-35% of our budget is being 100% effective (leads to a sale) while the remaining budget has a 33% of chances of convert to a sale at a later time. Again, this is a very simplified model, you may find out that working with real numbers adds an exponential value of difficulty *

The Machine Gun Model

So we got to the dreaded “machine gun” model. Here, our soldier will use a really high number of bullets to get the same results than the sniper. Take a look at the graphic below:

Here we can see that the amount of resources used to achieve the same results increases in an exponential rate. In the basic example, we have a 360 degrees market with 8 targets, so, to our machine gun soldier it will take 45 bullets to get a 100% target (sale). In other words, the budget will increase 45 times for 100% effective sales. You’ll also notice that brand awareness is a little higher in this model, and that’s why companies going for wide range branding won’t care much about niche refining (this is a very common usage in mainstream marketing, however, in my humble opinion it’s a waste of energy and resources in online adult marketing).

In short, it’s common sense to use the Sniper Model. However, most people uses the Machine Gun Model. Why? Well, because unless you’re part of a certain community or a very knowledgeable person about a certain market niche, the sniper model requires a lot of specialization and the use of every single resource available to refine data. Of course, data research is the most important one, but technology usage, smart design elements, competition status research, innovation, economics and so on are factors to keep in mind. As an example, I honestly doubt any designer could compete with me in certain niches (ie pornstars, glamour and some fetishes) when it comes to results. But, what if I need to do a design for an amateur mature site? Well, 99.99% of designers will say “Yes, I can do that” and come with the same cookie cutter design, just changing the subject. We all know it’s like that, and I can’t blame anyone on that, there’s not a lot of work for designers, and at the same time, some site owners aren’t well informed on the subject as well, so it all comes to a vicious circle.

However, information is power, and I’ve the knowledge to make things different, so why not use it? I’ve seen/heard many times the usual mottos “marketing is useless”, “you don’t need a designer to do a site, do it yourself because you know the niche better” (I gotta admit this is the funniest one, out of tens of thousands of adult webmasters I can count with my hands the number of people capable of doing this with decent results), “I’ve been doing this for years” (yet most old adult companies are closing doors) and so on and so on. And so on. I won’t say anyone is wrong, to each his/her own. However, if everybody else is failing, why keep following the path of failure?

So yes, information is power, and although it’s very uncommon to see someone providing secrets to competition, well, here here are some useful techniques.

Some real life techniques

Marketing Effectiveness Research

First define a budget, no matter how small it is as long as you can do something with it, you can grow it later. Isolate this marketing action by using specific custom campaigns so data isn’t contaminated. Now, choose 2 to 4 marketing models (ie banners, text links, galleries, different traffic sources, same marketing materials with different selling claims/colors and/or any combination of these) on a strict time frame, for example 1 month. Try to keep traffic levels to be consistent throughout the whole time frame, for example, if you buy 90,000 hits try to make those hits come at 3000 per day.

After the time frame ends, you’ll have some amount of data, including traffic (one column per each source), CTR, sales, time frame and so on. Using Excel, input all the data per day (1 row per day) including amount of traffic, amount of traffic to sponsors (or to join page if you have a paysite), amount of sales and a last column should include cost of traffic (if you didn’t buy traffic, use the average value from traffic providers for the same kind of traffic you have). Make a graph with this data so you can see the results at first sight. Do this for each marketing model.

Now, if you kept the traffic flow at the same level, the last column should give you a somehow flat line, but chances are you’ll see several hills and valleys in the other columns. Quite possibly, you’ll see differences depending on day of the week and day of the month. So now, let’s go to data analysis:

a) first thing first, subtract  your initial budget from the total amount of money derived from that campaign. If you get a negative number or a little positive value… well, you’re doing everything wrong, so time to rethink your strategy.

b) take your total number of sales and divide it by 30 (or any time frame you decided to go with) to get the average sales per day. Add it as a new column to your graph, repeating the same value on each row. Now you’ll have a flat line. Check out every spot where the value is above the line, and try to find something in common. For example: are those spots on weekends? at start of month? coming from some traffic source? is there a daytime/nighttime difference? (most probably, yes). Again, do it for each marketing model.

c) now compare the marketing models themselves and you’ll see a lot of information that will help you to diagnose which are your strengths and weaknesses.  Thus, you may end with something like “best results came from Tour A with TGP traffic on Wed around 6PM and 3 AM with geo IP US-DE-UK”

so in this example, it’s very easy to see that all tour versions converts better on Wed to Fri  of Week 1, between 6PM and 6 AM (day/4 means we have divided the day in 4 equal parts, but you can do it with 3, 4, 6, 8, etc. I like 6 myself), with geo IP us-de-uk. We also see tour A converts great on TGP, while tour B seems to be built for SE. Let’s say our original budget was $3000 using 3 tours and 3 different traffic sources all equally divided, thus we invested (or wasted) $66.66 per sale, and common sense tells us to divide our budget in a more efficient way since most traffic fell out of effective range. Now it will look something like this:

Model 1: Tour A, TGP traffic, 3 days a week (Wed to Fri), first week of the month, 6 hours a day (6PM to 12PM), US-UK traffic

Budget explanation: when we started, we had 3 marketing models with $0.463 per hour per traffic source investment ($1000 divided by 30 days divided by 24 hours divided by 3 traffic sources).  Now we’re weeding out 2 traffic sources, 4 days, 3 weeks, 18 hours and we realize that each sale in this targeted model would cost us only $8.33 . See the maths below:

0.463 (hourly investment) x 1 (traffic source) x 3 (days) x 1 (week) x 6 (hours)

so Model 1 real cost per sale is 12.49% of what it was.**

Model 2: Tour B,  SE traffic, 4 days a week (Wed to Mon), first week of the month, 12 hours a day (6PM to 6AM), DE-UK traffic

Now we’re weeding out 2 traffic sources, 3 days, 3 weeks, 12 hours and we realize that each sale in this targeted model would cost us only $22.22 (33.33% of what it really was) . See the maths below:

0.463 (hourly investment) x 1 (traffic source) x 4 (days) x 1 (week) x 12 (hours)

Therefore, although we could use a budget redistribution of $1875 for Model 1 and $1330 for Model 2 (or $1750 and $1250 for budget correction ), we still don’t have a tendency, so we’ll give tour C another oportunity, only that the budget will look more or less like this:

Model 1: $1700

Model 2: $850

Model 3:$450

After this time frame, we repeat the analysis, only that instead we’ll add a “previous time frame growth” column (in %): anything over 0 will tell you we’re on the right path. Any negative number will require further adjustments. Either way, in no more than 5 or 6 months you’ll probably have a defined tendency and a pretty targeted niche

Next Week: Our Sniper Tumbles! (and even more techniques) Updated: New post is on!

*if you’re interested in going deep further into statistical analysis and research, you should use Variance or Standard Deviation to measure the dispersion values

** according to the table, you may say that I’m leaving out the 19 sales, however in this example we’re working on time frame trends and the cost isn’t per sale, but for the best hour spot