Using keyword research for concept validation

Today I want to write about using basic SEO keyword research techniques to validate my early ideas about Eat Local Ballarat and get a sense of whether it was a workable opportunity.

First, the background: I’m planning to create a comprehensive, informative, up-to-date site on the subject of local food in the Ballarat region.  To meet its goals it needs to perform well in search engine results, build a strong following on social media, and generate plenty of visitors and buzz.

But how can I tell whether there’s any hope if it doing those things? What I need is concept validation: an early test that will show whether the concept is viable before I throw too much effort into it.

Concept validation for early-stage content sites

If this were a traditional business where I wanted to sell widgets, my first step would be to find a handful of potential widget-buyers and see if I could sell my widgets to them.  For a content-based website, the equivalent would be to just start blogging and see if it gets any traffic and engagement.

The problem is, it can take six months to start seeing a payoff in terms of search engine traffic.  Building an engaged community generally doesn’t happen overnight, either, as it can take a while for the network effect to ramp up.  What I needed was something that I could do in a day or two, before diving in too deep.

I figured there were three general elements that, together, could indicate this idea had legs:

venn diagram showing the three factors used for concept validation

 

First, I want to make sure that people are interested in the broad subject matter: eating locally.  Second, I want to look for existing food websites around Ballarat (if any) and see whether there’s a lack of quality local information on those topics.  Finally, I want to find successful sites in other cities/regions but covering the same niche, and see how they’re doing – this will give me an idea of what might be achievable.

Gauging interest in a topic

The first step was to gauge whether people are interested in local food and related topics, and exactly how interested they are in it.  I used two tools for this, both from Google and both free: Trends and Keyword Planner. (Note that you need an AdWords account to use Keyword Planner — you can sign up, create a campaign, then immediately pause it in order to use it for free.)

I started by making a list of terms related to my subject area, like:

  • local food
  • locavore
  • sustainable food
  • seasonal food
  • farmers market
  • community garden
  • free range eggs

(My list was longer – this is just a sample.)

My first stop was Google Trends, where I fed in a few of these terms to get a sense of their overall popularity.  Trends tells you whether the topic is getting more or less popular over time.  If you feed in multiple terms together, you can see their relative interest levels on the same chart.

screenshot of google trends chart for local food and locavore
“Local food” vs “Locavore” on Google Trends. Local food has more traffic, but is a vaguer search term; locavore is more likely to indicate well-qualified traffic.

Some search terms show distinct seasonality, like this one for “farmers market” which seems to peak in July every year.

screenshot of farmers market trends showing strong seasonal spikes
Trends for “farmers market” globally, clearly peaking each July.

This is a hint we’re mostly seeing northern-hemisphere searches, as farmers markets in the colder parts of North America often don’t run in the winter (our summer). However, we can use Trends’ “Regional Interest” data to just view searches in Australia, or even Victoria, and get a better sense of the local interest.

Screenshot showing steady increase in searches for farmers markets in Victoria
Trends for “farmers markets” in Victoria show that it doesn’t have the seasonality that global trends showed, and also that interest has been steadily growing for some years.

After the farmers market experience, just out of interest I went looking for terms that showed seasonality within Australian searches, and found “butcher“, which has clear peaks in December every year.  I’ll definitely be making a note to write a feature on local butchers in the lead-up to Christmas!

The good news is, for most of the terms I was interested in, there was an overall upward trend going on in Australia.  If I’d seen miniscule volumes or downward trends, this would have been a good time to ditch the whole idea, but luckily that won’t be necessary just yet. (On the other hand, if you’re looking to start a One Direction fan website, I think you’ve missed your big chance.)

Finding the most popular related terms

Now, Google Trends won’t actually tell you how many people searched for each phrase, only their relative popularity over time, but Keyword Planner can help us get more specific figures.

My next step was to put each of my terms into Keyword Planner, setting the location to Australia.  Initially I tried setting it to Ballarat, but the numbers were so small as to be useless, so I broadened my scope.  I’m assuming that Ballarat people’s interests and search habits are not too different from those across the rest of the country!

The results from Keyword Planner give you several useful pieces of information, but at the moment I’ll focus on just two: average monthly searches, and related keywords.  Below are some of the related keywords for “locavore”.

Screenshot of Keyword Planner's results for the term locavore, showing terms such as "eat local" and "local produce" along with their monthly search volumes.
Keyword planner results for “locavore”.

As I found new, related terms, I added them to a spreadsheet, along with their monthly search volume.  I kept building out from my initial terms until I had expanded my keyword list from around a dozen terms to well over a hundred.  Here’s a small selection:

Screenshot showing keyword research and monthly search volume for terms related to organic food.
A small sample from my keyword research spreadsheet.

Absence of local coverage

I sorted my keyword spreadsheet by the search volume, then starting with the most popular I fed each one into Google along with the word “Ballarat”. For instance, “farmers markets” would become “ballarat farmers markets”.

For each search engine results page (SERP), I reviewed:

  • Were there relevant matches for the search term on the first page?
  • Did the results cover the topic comprehensively? (based on my own knowledge)
  • Were the results current and up to date?
  • Were the results from local websites? (vs. national or international ones)
  • Did the pages appear to have expertise, authority and trustworthiness? (E.A.T. – these factors are what Google is looking for)

From these I developed a rough ranking of local coverage, simply a number from 0-3, and added a spreadsheet column for it.  My rankings were:

  • 0 = No relevant matches
  • 1 = Somewhat relevant matches, but they don’t cover the topic well
  • 2 = You can find decent information somewhere on the first page of the SERP
  • 3 = The first page of the SERP has one or more high quality sites with current information that comprehensively covers the topic
Search results for ballarat locavore
Google search results for “ballarat locavore” show pages from 2008 and 2013 (but it turns out the 2013 page actually has content from 2008 when you click through.) I’d give this a score of 1, since the information isn’t comprehensive and is so old.

Tip: if you’re doing lots of searches like this, use a “Private Window” (Firefox) or “Incognito Window” (Chrome) so that Google doesn’t apply search personalisation. If you’re often searching for related topics, it might helpfully try to rearrange the results based on your previous activity. You want it to act as if you’ve never searched for this before.

I added three more columns to my spreadsheet:

  • Quality of local search engine results (0-3, as described above)
  • Best local site (based on my own judgement), into which I pasted the web address of the page in question
  • Domain Authority (DA) of the best local site.  If there was no local site, I left it blank.
  • Facebook and Twitter follower counts for the best site

I knew that if I saw lots of “3s” for coverage, along with lots of high DAs and social media followings, there would be stiff competition in the area.  Conversely, if I saw lots of low coverage scores and DAs, I’d know nobody was covering this topic well yet.

The results were mixed: some topics had good coverage (local wineries and fruit trees among them) but most topics were a bit disappointing. You could often find relevant information if you went digging through a few links, but not much that was comprehensive and up to date.

Median DA of the “best” local websites in the results (excluding a couple of mostly-non-food outliers like the local newspaper and the City of Ballarat websites) was 16, and median social followers (Facebook and Twitter combined) was around 400.

Are comparable websites successful?

I wanted to find out whether people in other cities had done something similar to what I’m planning, and how it was working out for them.  To do this, I defined what “similar” was, and came up with: a local food website (or a sustainable living site with lots of food-related content) dedicated to a geographic region with a similar population to ours, which I defined as being up to around 200,000 people (Ballarat has 100,000).

Next, I started feeding my most promising keywords into Google and looking for websites that might meet my “similar” criteria.  I used a lot of my own knowledge to guess at terms that might show up relevant sites, and focused on places where I’d already heard they had a strong local food culture, such as California, the Pacific Northwest of the US, the southwest of England, as well as Australia’s best known food regions.

As I found comparable websites, I made a note of them in another spreadsheet tab, along with their Domain Authority and Facebook and Twitter follower figures.  I also kept some rough notes of any ideas that the websites sparked, which I might be able to adapt for Eat Local Ballarat.

What I was looking for here was websites with relatively high DAs and good engagement (keeping in mind their niche topic matter and regional geography), and I found them.

  • Transition Town Totnes (a UK based local sustainability group who do a lot of food stuff) has a DA of 40 and over 6000 social followers, and this is for a town of just 8,600 people!
  • The US-based “Edible” magazines provide local food coverage for various cities.  Examples of smaller cities/regions include:
  • Daylesford Macedon Produce (a region just north and east of here, with a few towns of around 2,000 people each) has a DA of 29 and 1800 social followers.
screenshot of the daylesford macedon produce website
Daylesford Macedon Produce, serving an area just north and north-east of here, is doing a fabulous job.

On the whole, I found a fairly good range of sites with higher DAs and social media followings than the Ballarat sites I’d previously researched.  So it is possible to be pretty successful with a website about local food, in a regional area!

Putting it all together

To summarise the insights I’ve gained:

  • Interest in topic: Search terms related to local food are steady or growing in popularity in Australia. Some of the most popular terms get around 8,000 searches a month Australia-wide.  I now know which keywords are the most popular.
  • Absence of local coverage: When searching for the most popular terms along with the qualifier “Ballarat”, there aren’t many good (comprehensive, recent, expert, authoritative, trustworthy) local results. Answers are scattered and often include generic, non-local directory websites.  DAs are mostly in the teens.  Local websites covering this topic have social media followings in the hundreds to low thousands.
  • Success of comparable websites elsewhere: comparable local food websites in other regions can achieve DAs 30-40 range, and social media followings in the 2000-6000 range. Reviewing their technical setup and publishing practices, it looks like they could be doing even better if they wanted.

In conclusion: there’s a gap. We have an interest in the topic, the topic can be covered in such a way that it will get good traction, but local websites aren’t currently achieving that.

Based on this, and my previous work on finding the best existing Ballarat websites, my guess is that a focused local food site for Ballarat, with good attention to SEO and other inbound marketing practices, should be able to achieve a DA of at least 30 and a couple of thousand social followers.  That would put it among Ballarat’s top sites, and sounds like a worthwhile project to take on.

Concept: validated!

2 thoughts on “Using keyword research for concept validation

  1. WOW, Alex. That is some amazing research. So in depth. I have to admit, I’m not that crash hot at all this sort of stuff….it’s the techy side of things that sort of makes my eyes glaze over. BUT, I know I should pay attention because SEO does pay off.

Leave a Reply