The first post on my twitter account was "what is this and why would anyone want to use it?"
It was definitely not "this is precisely the thing that I had been hoping for 10 years somebody would code up."
If you had tried to estimate the number of potential twitter users by counting up the number of people in the world who were search for a service that would publish their SMS-length plain text messages on a web site, you would have gotten nowhere meaningful.
It's an excellent article overall. I didn't realize the search count include the adSense matches (#4). Searches from Google's partners should still be counted though; presumably the end users search the term on the partners' sites.
That's because #4 is wrong. Looking at my own campaigns, I get WAY more impressions from the content network than Google estimates in the keyword tool.
Also, evaluating your target market based on a single keyword seems pretty silly and meaningless. What about related or synonymous keywords? What about other traffic sources like banner ad media buys?
#4 is not wrong. It's a rule of thumb. It's not 100% correct all the time, but it's the best estimating approach we have. If you have a better one I am all ears.
>> What about related or synonymous keywords?
This post does not go into detail on the full approach for measuring a market. That's detailed in my book and way too long to include in a blog post.
This post simply talks about the 4 most common mistakes people make when doing this kind of research. Nowhere did I indicate your entire market research depends on this single keyword. That would be stupid.
>>What about other traffic sources like banner ad media buys?
There are at least 50 other traffic sources: Twitter, blogging, podcasting, other search engines, magazine ads, TV ads, radio ads, cold calls, direct mail, Facebook pages, Facebook ads...I could go on and on.
None of them offer data that allows you to easily perform market research. Also, unless you are funded many of them are completely out of reach.
Again, Google is not the only source of traffic but it's an easy way to measure market demand.
That rule of thumb is based on false assumptions. Here's actual data for September from a campaign I'm running on both search and content. I took the exact same keywords from my search campaign and copied them into a content network campaign:
Campaign:Clicks:Impressions
Search(+partners):9,738:244,005
Content Network:32,648:7,202,056
Adwords tool estimates about 200,000 local searches for my main keywords.
The Adwords Keyword Tool DOES NOT report impressions from the Content Network as a mesaure of search volume, that is just a fact.
Your method is fundamentally flawed. This process will tell you absolutely _nothing_ useful about the size of your market.
I hope your book does not have any similar errors.
I think the author means "target audience size" or "nr of potential customers" instead of "market size". A market size is usually expressed in money.
For many startups --as they often operate on new markets-- the size of the "target audience" is probably more interesting. In the markets the 'size' (in money) is also quite hard to measure. But in developed markets the 'size' is probably more interesting.
You're correct - I did mean the number of potential customers rather than the total amount of money.
I haven't heard that distinction, but I'll consider using your suggested alternatives in the future.
At the same time, I wonder if using a longer phrase like the "number of potential customers" would actually make the article less readable. I imagine most people knew what I meant by the term "market size" based on the context.
"The Google AdWord Keyword Tool doesn’t show you how many people search Google for this term each month. It shows you how many AdWords ad impressions are available. And that includes Google’s search partners, content partners, and the entire AdSense content network."
Can you cite a source for this? Why would it say "searches" if it counts AdSense impressions?
The more likely point is that the number 1 ad or search result doesn't get 100% of search clicks, they are distributed among all results on the first page.
Just because a user might search for the terms every day doesn't mean it's necessarily even possible for them to convert more than once (like for a subscription service). On top of that a lot of searches are really just content-addressable bookmark lookups.
"searches" is some multiple of "customers searching"
It was definitely not "this is precisely the thing that I had been hoping for 10 years somebody would code up."
If you had tried to estimate the number of potential twitter users by counting up the number of people in the world who were search for a service that would publish their SMS-length plain text messages on a web site, you would have gotten nowhere meaningful.