HN Explorer

Ask HN: How accurate is Google Adword's Traffic Estimator?

21 points by ashishk ↗ HN
I'm about to launch a startup, and plan to acquire users via search ads.

I'm using Google's Traffic Estimator to gauge how many users I can acquire in a month, but I'm not sure how accurate it is. Also, I'm not sure if it includes Global or US traffic.

Does anyone have experience with this?

Here's an example of what I'm looking at: https://adwords.google.com/select/TrafficEstimatorSandbox?save=save&keywords=hacker%20news%0A%22hacker%20news%22%0A[hacker%20news]%0Ahacker%20newss%0A%22hacker%20newss%22%0A[hacker%20newss]&currency=USD&language=en

19 comments

[ 7.6 ms ] story [ 352 ms ] thread
zaidf
Great question. I'm wondering the same about Facebook.
ashishk
Facebook will tell you the # of people you're targeting, but not the expected clicks. They might not have enough data to estimate clicks. Google indexed my site, and probably has more comparable data than FB does.
zaidf
Facebook does tell you the estimated clicks.

I'd like to know how accurate the estimates are.

ashishk
Really? I see a recommended bid, but not estimated # of clicks.

Let me know and I can let you know how accurate it has been for me (been running test ads for 2 months).

zaidf
You should see "Estimate: 74 clicks per day" at the bottom right above the Create button.

This is ONLY if you do PPC. If you do CPM, you'll see impressions(which is probably what you are doing).

If you are doing CPM, mind sharing your click %? I'll do the same when I start my campaign and hopefully we can learn together!

agbell
Here are a couple tips:

- Use the keyword tool instead of traffic estimator: https://adwords.google.com/select/KeywordToolExternal

- Look at the difference between exact, phrase and broad matches. (On Keyword tool)

- Rumor has it Google changes the broadness of broad match to suit their needs. I have heard of "Bed Frame" broad match showing ads for "mattress" Therefore the broad match number is probably higher than reality.

- The country settings seems mainly useless.

- The numbers returned are rounded to certain fixed values which seem to spread out in a logarithmic scale. So the higher the estimate, the less accurate.

transburgh
I have been doing paid search for a couple of years and have had different results when comparing the Traffic Estimator to Adwords results:

The most common outcome I find is the estimator tool shows more traffic and a lower cost per click than actual adwords results. I have even had the cost per click off by 40% more than the estimator tool.

The problem with the tool is there are too many variables that the tool can not account for. Other than cost and keyword, you have to deal with the quality score that looks at the ad you write, the landing page the ad is pointing to and click-through-rate (ctr).

In the SEM industry, the rule of thumb is to take a small amount of money and test paid search. It is the only way to get a good idea on the environment you will be competing in.

chasingsparks
I had a similar experience. I naively used the Google Traffic Estimator tool very early on in the planning process of my startup. Unfortunately, I gave a strong weight to these estimates -- which turned out to be very, very wrong (I was using it to find CPC and traffic for major metro areas).

Lesson learned. Get lots of estimates and be as conservative as possible.

EinhornIsFinkle
1. You select the area (country, city, whatever) you want to target.

2. Take the numbers with a grain of salt. They're good for directional or relative comparisons (same KW in different parts of the world, or KW vs KW) but by no means present an absolute measure of demand.

effektz
I've found that most of the time, the Google Traffic Estimate usually OVERESTIMATES how much traffic each keyword get. For example, I have the #2 google spot for a food keyword. I get about 1,500 visitors a month from Google for that keyword, but the traffic estimate tool estimates "74,000" searches for that keyword a month. So if 74,000 is really accurate, only 1,500 are clicking the second result
keltex
2% clickthough rate is a little low, but not out of the realm of possibility.
drp
Depending on the keyword and the title/description of each result, those numbers entirely possible. I've seen exponential traffic differences between 1st and 3rd results for a few keywords, so I'd have to guess that the 2nd result would behave similarly.

You also have to consider the searches that result in no clicks or a modification of terms.

It would be very interesting to see actual click through rates for each spot in a few representative sample search results.

mcdowall
MarketSamurai should give you a wealth of information about your intended keyword.

(Disc - no affiliation to the product, i just use it a lot!)

weaksauce
I couldn't find a price on their site. How much is it? Also, does it work without needing to buy a subscription?
ddemchuk
MS is $97 one time, no subscription. it really is one the best search data tools I've ever used
weaksauce
Thanks! Do you know if they are using data from their own servers or if they are just pulling it from the web on the fly?
ddemchuk
I believe nearly all of their data is just pulled from the big 3 search engines, especially Google. They also grab page ranks, backlinks from yahoo, adwords data, and microsoft oci data. It's not proprietary data, it's just presented really well and saves a hell of a lot of time
JoelPM
A year ago it wasn't extremely accurate. When I left there was a project underway to significantly improve it, but I'm not sure if that happened.
dcnstrct
Not that accurate -- a better method is to use the Google keyword tool and numbers you have internally on response rates. You will get a more accurate range that way.

Of course, relevancy of the ad to the keyword is the critical factor. If your ad or service is not appropriate to the keyword then you cannot expect a good response rate.