I've become a bit obsessed with how Google counts results. You may think it's easy (1, 2, 3… 10), but add in 7-result SERPs and blended local results, and counting to 10 is no longer a Kindergarten-level achievement. Pictures speak louder than words on this one, so let's look at an example. Here's a localized but de-personalized SERP for "orthodontist" — I've stripped out everything but titles, display URLs, and pins, to make it easier to parse:

The two sets of numbers on the left represent the two ways I think most rational people without local SEO expertise would count these results — it's either six "pure" organic results, or 13 total results. The problem is that almost all page-1 Google SERPs have either seven or 10 organic results. So, there's a third interpretation — this is a 10-result SERP, but some of the local 7-pack (in this case, some = four results) must be "blended" results. In other words, the local pack contains both truly local results and organic results that are being treated as local.

Hacking the start= parameter

So, how do we figure out which four are blended? You're probably familiar with Google's "start=" URL parameter. Even if you don't ever enter it manually, you use it all the time — it's what separates Google's search result pages. So, if a basic query looks like this: 

<a href="https://www.google.com/search?q=orthodontist">https://www.google.com/search?q=orthodontist</a>

...then the query to reach page two of results looks like this: 

<a href="https://www.google.com/search?q=orthodontist&start=10">https://www.google.com/search?q=orthodontist&start=10</a>

It turns out that "start=10" is a bit of a cheat – it always means "jump to page 2" even if page one is a 7-result SERP. Like most coders, Google also thinks in terms of starting at zero, so for a traditional SERP "10" actually means the 11th result (page one is results 0-9).

Here's where it gets interesting. What if you change the "start=" parameter to be something other than a multiple of ten? Turns out, it works just fine, and it gives you a stripped-down organic result page starting with the absolute position specified. In other words, setting "start=9" gives you a page with no local results that begins at the 10th organic ranking.

Counting backward to destiny

Ok, "destiny" may be a bit over the top. It turns out that you can effectively use this technique to count backward and determine the "true" organic results, as if the local pack had never appeared. You can skip straight to "start=1", which shows the 2nd ranking forward ("start=0" is Googlese for "start from the beginning", so you have to make some assumptions about the #1 spot).

Using this trick ("&start=1") for my "orthodontist" query at the beginning of the post, I ended up with these results:

Since we're starting with #2, this page actually represents organic results 2-11. It's a little odd, but hopefully that all makes sense. So, why am I torturing you with these mental gymnastics?

Putting it all together

If we match up the URLs in the second list with our original SERP, we can determine not only which results were blended, but also what order they would've appeared in without the influence of the local 7-pack. It looks something like this (organic results are in green, local are counted with "L" in the number):

In this case, the first four local results in the pack are the blended results, but the 4th result is actually #9 in the original organic results. Like the old indented results, the local pack pulls any organic result that gets promoted up (to keep the pack contiguous), so in this case #9 is actually outranking the original #7 and #8.

Finding hidden opportunities

This may seem like an academic exercise, but a conversation with local SEO expert Mike Blumenthal helped me see the strategic importance. Understanding how local and organic blend in the SERP above, for example, tells us a couple of things. Google back-filled the 7-pack with three purely local results, indicating an opportunity for sites that might be weak on organic ranking factors but are decently optimized for local.

There's also a potential opportunity for some of the lower-ranking organic results to get promoted above other organic results by improving their local ranking factors. For example, #10 could jump above #7 and #8 (using the organic counting method) with some solid local SEO efforts. In the overall SERP, #10 could jump just behind #9, effectively gaining five spots.

Effectively, there are two algorithms in play here, and they overlap. Local is no longer a purely independent consideration, and "blending" is a dynamic process that potentially opens up new opportunities. We're going to see this with more and more "verticals," including Knowledge Graph — these features will start to cross over into organic results and modify them with specialized sub-algorithms. Being visible in these SERPs will require an understanding of how all of the pieces fit together.