Keyword analysis on a startup’s budget

We’re working on identifying terms people use to search for stuff related to what we do. The aim is to help us understand what’s important to our customers and what words we should be using to talk to them with. This analysis is often the job of specialist agencies but how do you do it when you need to do things as cheaply as possible?

Initial terms

First things first, what are the words we think we need to be using? These are going to be the fundamental things you want to offer or support. These might be the word or phrases you expect to see in tag lines and short descriptions.

  1. Data Strategy
  2. AI Strategy
  3. Data Science
  4. DataOps

We can start gauging how much competition there is in search and paid advertising around the words we want to use with Ubersuggest

Ubersuggest results for AI Strategy — 63% chance of ranking in the top 20 results in the US

However, if we only worked on just a few terms our website and blog posts would be exceedingly boring and repetitive and we’d be putting all our metaphorical eggs in just a few baskets. We need to diversify what we talk about so we need to find a set of related topics.

Related topics

Ubersuggest does have a keyword ideas section but I found that BuzzSumo provided a neat topic tool that helped identify clusters of related topics aka keywords. BuzzSumo have a free trial which you can use to get started and with the content suggesters etc. I think it’ll be a tool I come back to in order to help us identify meaningful content to share.

BuzzSumo Topic Explorer for DataOps

With a OneNote filled with candidate keywords, the next step is to identify which ones would be good to think about or talk about sooner rather than later.

Detailed keyword analysis

To be able to do the next stage I was interested in a solution that would give me an API or a bulk-upload interface for free or cheap. Google Ads has a Keyword Planner but having to create a campaign before I could even get started wasn’t fun. Another recommendation I’d had for keyword analysis was and I noticed they had a $7 trial and an API that looked well documented. The API didn’t seem to cover keywords but it did have a bulk upload capability. I copied and pasted the lists out of OneNote into a .txt file and uploaded it into ahrefs. I could then export these as a CSV which enables analysis in Excel or some other solution.²

The results for our 400+ candidate key terms

Working out which keywords to prioritise is based on things like how many searches there are, how much competition there is, and how many people actually click through. A rudimentary starting point is to look at a few of the key metrics and rank words based on those.

Selecting the top 5 most difficult search terms and their monthly search volumes. Uses R and the tidyverse 📦.

Some things I’m looking at include understanding the words that’ll be toughest to succeed at ranking well. Looking at the list, Artificial Intelligence is one of the words we’d want to rank well on so it looks like we’ll need to put in a lot of effort to achieve a high ranking for the term if we go after it.

All keywords with a difficulty lower than the mean plus a standard deviation and a higher than average search volume. Uses R and the tidyverse 📦.

Words that we might want to do well at are words with relatively low difficulties but reasonably high volumes of searches. This is part of marketing to the long tail, the more niche searches that are less contested and easier to rank higher and more consistently in.

All keywords that have a lower than average cost per click but a higher than average search volume. Uses R and the tidyverse 📦.

As we’re on a budget (having spent a whopping $7 so far!) if we want to start doing pay-per-click (PPC) advertising (e.g. Google AdWords) we’ll want to know things like which keywords get a decent amount of searches but don’t cost a lot to bid on.

Searching for all keywords where we can rank well (helps conversion) but still get a low cost per click on a relatively large number of searches, and getting an estimated cost for a keyword campaign based on 3% of searches resulting in an ad click. Uses R and the tidyverse 📦.

Ideally, we would get the information about the percentage of clicks for a keyword that are PPC such that we could start estimating the cost to run a campaign but we don’t so saying 3% based on a sample on the webpage, we can put together a number of filters for keywords and estimate the cost winning 100% of the clicks for our cheapest keywords.

These sorts of views of our keywords mean that we can start prioritising which keywords to target for content and advertising. We’re able to understand how popular different terms are and how people are engaging with the results. Not bad for a total of $7!

What next?

Well, I’m definitely not an SEO or PPC expert but I do know an important next step is to incorporate the keywords that competitors are targeting. The selection of keywords overall will probably merit more human curation for sense, relative to the goals of the company.

Finally, once there’s a more finalised and prioritised list of keywords, the next level of detail in a bootstrapped marketing plan is to start developing content and ad copy for keywords.

¹ Thanks to Blythe Morrow, of Paper Sword B2B for this tip.

² I’m using the programming language R which is great for data science.