Only by measuring, can we determine whether our advertising and marketing efforts are paying off. My approach tends to be ‘measure as much as possible’, because it is in aggregate that much data begins revealing insights.
In today’s post, I am going to be tackling a classic marketing measurement: share of voice. Share of voice is a way of measuring an amount of advertising presence or strength for a particular brand or product, or essentially: how much is our brand or product talked about compared to our competitors? The beauty of the internet is that we have a way of measuring how much we are being talked about, or what percentage of our audience’s attention we have.
The traditional view of share of voice
Traditionally, share of voice is measured like this (this is in the case of a brand - replace ‘brand’ for ‘product’ if that’s what you’re dealing with):
Share of Voice = Brand Advertising (£/$ or #) / Total Market Advertising (£/$ or #)
Advertising might be based on spend, or other metrics such as impressions or gross rating points (GRPs).
It is impossible to get a completely accurate figure for share of voice. In the case of advertising spend above, you can only estimate how much your competitors are paying for advertising, and you wouldn’t be able to get any more accurate idea other than by getting the figures from the advertiser or ad platform themselves (something that is not going to happen).
I am going to give this classic KPI a digital marketing spin a look at how you can calculate it for organic search. There are limitations with this method, holding back how accurate your share of voice figures are, but I will try and be as thorough as possible (without the use of very expensive software).
How can you calculate share of voice for organic search?
This isn’t the only method around, but it’s fairly simple to do. Essentially, we want to look at our brands occurrence in the SERPs for a group of keywords we are interested in, compared to that of our competitors. While share of voice is traditionally based on impressions, we're going to look at clicks instead, because if one result is getting more clicks, it's almost certainly getting more attention.
This job is made a whole lot quicker and easier with paid tools, but I want to show you the manual method first so you know how it works, and in case you don’t have access to any tools to help you. If you want to find out more, Sycara is one tool that can help you calculate share of voice.
1. Choose your keywords
First of all, decide what keywords you want to calculate your share of voice for. I’m going to use the supermarket Tesco as an example here, and calculate share of voice for five keywords:
- online grocery shopping
- online food shopping
- food shopping online
- supermarket delivery
- grocery shopping online
You will want to choose a group of keywords within a common theme (based around a specific product or service), and realistically, you’ll want to use a larger list than this (which is one of the reasons why using a tool to automate some of this process helps).
2. Choose your competitors
This step can technically be missed, and you’ll still be able to calculate a share of voice figure. But, that share of voice figure won’t really mean much if you haven’t got anything to compare it to...
You’ll want to decide whom you are going to be comparing yourself against. For the purposes of this example, I am going to choose four competitors to Tesco (all British supermarkets):
- Sainsbury’s
- Asda
- Waitrose
- Ocado (Ocado actually do sell Waitrose products, but I will ignore that for now)
Make sure you know the URLs for the online properties of all of your competitors; a competitor might have more than one domain, or might use sub-domains.
3. Record the ranking(s) for each keyword
This is the time consuming step if you’re doing it manually. For each of the keywords you selected in step one, you need to record whether you rank for them, and if you do, in what position (be sure you’re not receiving personalised results by going Incognito with Chrome, or by using the Moz extension’s custom search feature). There is only point looking at the top 10 or top 20 results; the top 10 results receive almost all of the clicks, but being in the top 20 might mean the rankings for the website you’re measuring are going up. So in the case of Tesco:
|
Keyword Tesco |
Ranking #1 | Ranking #2 | Ranking #3 |
|
online grocery shopping online food shopping food shopping online supermarket delivery grocery shopping online |
7 3 2 8 8 |
8 9 |
Remember to record all positions you appear for - including multiple domains/sub-domains. Now, do this for each of the competitors too:
|
Keyword Sainsbury's |
Ranking #1 | Ranking #2 | Ranking #3 |
|
online grocery shopping online food shopping food shopping online supermarket delivery grocery shopping online |
2 8 8 5 |
11 9 9 11 |
|
Keyword Asada |
Ranking #1 | Ranking #2 | Ranking #3 |
|
online grocery shopping online food shopping food shopping online supermarket delivery grocery shopping online |
3 2 3 13 3 |
4 4 |
|
Keyword Waitrose |
Ranking #1 | Ranking #2 | Ranking #3 |
|
online grocery shopping online food shopping food shopping online supermarket delivery grocery shopping online |
5 5 5 4 6 |
6 6 5 7 |
6 |
|
Keyword Ocado |
Ranking #1 | Ranking #2 | Ranking #3 |
|
online grocery shopping online food shopping food shopping online supermarket delivery grocery shopping online |
6 4 4 2 2 |
Now, I highly recommend using a tool to automate this for you because it will be a long process when you’re using many keywords. A good place to look is https://my.webposition.com/ - you will be able to set up reports that do exactly what we’ve done above: find the rankings for a number of different domains for a number of different keywords. You can then export the results to a CSV and work on them from there. But, whether you use a tool or not, keep your results in a spreadsheet for the following steps.
4. Calculate share of voice for each keyword
The last step is to calculate the share of voice. There is a reason why we’re recording the positions of the keywords in the SERPs, not just whether they’re there or not: because different positions receive different numbers of clicks. If you had five keywords in the top 10, and so did your competitor, that doesn’t mean your share of voices are the same - it all depends on the rank. For this reason, we will calculate share of voice using average click through rates (CTRs) for SERP positions. There isn’t one single set of values for average CTR, but I’ll be using the results from a study by Optify, but feel free to use different values if you prefer (for example, the results from Dejan SEO).
| Rank | Average CTR (%) | Rank | Average CTR (%) |
|
1 2 3 4 5 6 7 8 9 10 |
36.4 |
11 |
2.6 1.5 1.3 1.1 1.2 1.2 1.4 1.3 1.4 1.4 |
For each single keyword, you need to add up the click through rates of as many positions as your (and your competitor’s) website received. So, if you ranked 3rd and 7th for a keyword, your total CTR for the keyword would be 9.5% + 3.8% = 13.3%. This is also your share of voice; you will be receiving 13.3% of the clicks for searches on that particular keyword, and therefore your share of voice is the same: 13.3%.
To do this quickly in your spreadsheet, you’ll want to have the average click through rate table in your document and then refer to those values using the VLOOKUP function. VLOOKUP, in Microsoft’s words, “searches for a value in the first column of a table array and returns a value in the same row from another column in the table array”. To put it more simply, look at the example below from my spreadsheet:
I want my spreadsheet to add up the average click through rates of all three of the rankings for each keyword. Tesco ranks in 7th and 8th place for the keyword “online grocery shopping”, so my three ranking values are 7, 8, and 0. I’ve entered 0 for no ranking to avoid complications with the Excel function - I’m not an Excel expert and there will be someone with a more elegant solution, but hey - this works! For the same reason, you’ll notice in my table for average CTRs that I’ve put a 0% for a 0 rank.
The share of voice Tesco has for “online grocery shopping” is 3.8% + 3.5% + 0% = 7.3%. The VLOOKUP function in my spreadsheet looks at the ranking position, looks in the CTR table, and returns the corresponding CTR % for the rank. It does this for all three rankings with three VLOOKUP functions added together.
The VLOOKUP function looks like this:
VLOOKUP(lookup_value,table_array,col_index_num,range_lookup)
lookup_value = the cell containing the keyword ranking. So for the first ranking Tesco has for “online grocery shopping”, that cell is B4.
table_array = this is the table the function is going to look in to get the CTR value for the lookup_value. In my case, this is the top left cell of the table (excluding the table header), H3, and the bottom right cell of the table, I23, so the table_array value is H3:I23. I have added $ into that value to make it $H$3:$I$23 so that those values are locked - when I copy and paste the function into all the other cells that need it, the table_array value will not automatically update itself and change.
col_index_num = is the column in the table array from which the matching value will be returned. In my case, I want the function to return a CTR rate value, which is in the second column of my table, so col_index_num is 2.
range_lookup = is the logical value that specifies whether an exact or approximate match should be found. FALSE returns an exact match and doesn’t require the table_array be sorted in any particular way - perfect for our example.
Now that we’ve gotten that out of the way, you should have a share of voice for each keyword, for you and for your competitors. This is the example for Tesco:
5. Calculate overall share of voice
Now we’re ready to calculate an overall share of voice for all keywords together. This is as simple as calculating the average share of voice, so for Tesco that is: (7.3 + 9.5 + 12.5 + 3.5 + 6.5) / 5 = 7.86%. That’s all there is to it!
6. Refining the calculation
The above figure for share of voice will give you decent idea about how your presence and rankings in the SERPs are stacking up against that of your competitors. But, it assumes that each keyword is worth as much as each other, when in fact they all have different search volumes.
If we want to have a more accurate idea of how much potential traffic we’re getting for a group of keywords compared to others, we need to account for the difference in monthly search volumes. Back to the spreadsheet then.
Add an extra table with the estimated monthly search volumes for the keywords you’re analysing.
Calculate the estimated search traffic you’ll be receiving for each keyword like this: keyword share of voice (as a percentage) x keyword monthly search volume = estimated monthly search traffic.
Then, to get your ‘weighted’ share of voice value (one that takes into account monthly search volumes), add together the estimated monthly traffic for each keyword, and express that as a percentage of the total monthly search volume of each keyword. So for Tesco this is ((116.8 + 152 + 110 + 11.2 + 20.8) / 4720) x 100 = 8.7%.
Share of voice is a great competitor analysis metric
So there you have it, the long and quite tedious journey towards calculating share of voice. To make your first steps a little easier, here is my spreadsheet for you to download and do with what you will:
Download my spreadsheet to get started.
You'll find that share of voice is a great metric for comparing one brand against another, and without gaining access to your competitor's analytics data, is one of the better ways for seeing how much traffic one site is receiving over another in search engines.
If you've got any questions, let me know in the comments. Want to offer your own way of calculating share of voice? Let me know too!
Great post. Peter. This is a pretty interesting way of reporting an important metric. I do think that the math is going to get much trickier especially for large sites that optimize for multiple locations.
Because Google is personalizing search results in so many ways now (location, device, search history, etc.), it's going to be more and more difficult to rely on reporting like this because there are more and more variables to deal with. For example, you may have a different ranking for a single keyword on different devices. And those rankings may change based on where the search is done. And those search results may be different based on what that individual searcher has searched before.
I love this kind of reporting and the insights we gain from them. Unfortunately I think Google will continue to make it harder to do these (at least manually). Luckily SEOs have awesome people on their side like those at MOZ to make powerful tools to deal with all of these changes.
Thanks Mike. You're right, the calculation becomes much more difficult with larger sites, and it's here that there needs to be a large amount of automation. I might revisit the idea for a future post, looking at a website where we're analysing perhaps 40-50 keywords in a particular niche.
Can't wait to see it. I currently work in-house and I monitor multiple keywords in multiple markets on multiple devices. The math gets very sticky and would surly defeat a lesser man. But, if anyone can navigate those numbers, I'm confident it is you.
Looking forward to your next post, good sir.
I am trying really hard and these are great metrics, but I wish that mattered. I am still looking at the daily visitor count and maybe that's the whole problem.
Hi Peter,
It's really interesting method for comparison between two different brands without the access of competitor's analytic data. Thank you so much peter for showing up this tricky method. its very simple. Waiting for your next post with helpful information.
Peter Meinertzhagen well done! You had done the great job for sharing such an amazing piece of information regarding the voice contribution. Your described calculation well elaborates the share of voice for organic search.
Great job @Peter! However, I am also wondering about the numbers in your calculation. Why don't you add all the competitors whose keywords appearing in 20 top results in SERP?
Hi, thanks, glad you enjoyed it. The simple reason was for the purposes of simplicity for the calculation. Calculating that many competitors manually would be a long task and one best suited to an external tool (e.g. my webposition) where you can automate some or all of the process. Some expensive tools can do the whole thing for you. Hope that helps!
Thank you for this great little sheet and discussion. I immediately took your file and went "Advanced Excel" on it to create something I could scale up.
I've found almost immediate usefulness for this because our agency is trying to win back a client that cancelled over a year ago. Clearly a dramatic change in SOV means that high volume words have moved in or out of top 3 - I'm already looking at a drop of over 6% with them. Bullseye!
Sorry to be a party pooper, but while I hate the Share of Voice metric and would not champion it's use, this isn't the way to calculate the "true" Share of Voice.
Share of Voice (in the world of organic search) is the calculation of how far reaching your website is across all of your product base, which can then be compared to your competitors.
Therefore only calculating it based on a small set of terms is pointless. Even calculating it based on 1,000 terms is pointless. You need to calculate it on 10,000's or even 100,000's of terms to even begin to get a semi-useful number to make comparisons with.
For (a daft) example; If the department store Debenhams.com ranks 1st for "dresses", "hats", "watches" and "perfume", while Johnlewis.com ranks 10th for those terms, then you'd make the assumption that Debenhams has the greater share of voice in the fashion sector. But what about the millions of other keywords? It may be that those 4 terms are the only terms Debenhams actually rank for while John Lewis actually ranks 1st for everything else!
To use Share of Voice to the best of it's ability (and it's ability is not great in the first place!), you need to get as much of a holistic view of all the site's rankings as is possible. For this I'd strongly suggest SearchMetrics as you can export pretty much every single measurable ranking that a website has at the click of a button. You can also skip the VLOOKUP of CTRs with SearchMetrics as well as it calculates all of that for you...
...which will save you valuable time which you can use to ponder why you're wasting your time calculating the Share of Voice in the first place!
Hey Paul, cheers for your feedback.
I agree with you that in the example I've given, using only five keywords for SoV for a broad sector like online grocery shopping is way too narrow, and wouldn't give you an accurate or reliable result. In this post, I really wanted to show a basic example so people can see the principle involved, and then adapt it to their needs.
I disagree though that SoV is a pointless metric. If rankings are at all important to you (and they might not be) then SoV has the potential to be of some use. And not every category will require 1000s of terms to be useful. It all depends on what the website is.
There will be cases where you can track a fraction of those keywords, for you and your competitors, which will give a SoV calculation that is accurate enough to be of some use. This is of course needs to be judged on a case by case basis, but obviously leans towards less competitive niches. If you also use some AdWords, you can get a pretty good overview of the keywords you should be tracking in a particular niche.
I agree that wherever possible you should be trying to automate though. I wouldn't recommend using a fully manual method when using thousands of keywords! But you cannot understand how SoV works if you simply let a tool automate it all for you - I believe it's good to have a go yourself first.
Also, I disagree that this isn't true SoV. It is correct, only very limited and on a small scale.
I'd recommend people use this metric if they have a good way of reducing the amount of time it takes calculating, but not if it eats up a significant amount of their time each month. For this reason, I wanted to write a post in the future looking at how it can be automated as much as possible, for a large chunk of keywords.
Thanks Peter, I am just returning to calculating share of voice so your post couldn't have come at a better time.
I have to say it was nice in the days before (not provided) where you could sense check data against actual traffic received on a keyword-by-keyword basis, but I like this approach.
Hi Matt, you're very welcome. Yes, it would be nice to compare the results against actual traffic, but unfortunately that's no longer an option. Using AdWords in conjunction will give a good idea of the range of keywords you should be tracking though.
Great post, thanks for sharing! Market share value metrics seem to be often forgotten by marketers (due to the complexities involved in calculating them accurately I guess). We're just started working with some very cool software for our SEO clients in Australia which actually does all this and then analyses historical AdWords CPCs & conversion metrics to report a relative $ value market share metric. More on that here if anybody is interested and I'd love to hear from anybody who is using similar software for SEO market share reporting. Am looking forward to seeing technologies like these develop in the coming months & years!
Thanks Dave! I'd love to try out some software that automates a large chunk of this process. I mention a few examples in my post, and I might do a write up of them for a future post.
Will have to dust off the Excel Skills!
Indeed! I'm no expert in Excel and I'm sure there are ways in which my method can be improved, but it does the job.
Hi Peter
Great of you to share this metric. We have been using a similar model for the last couple of years (when "not provided" started showing up). It gives clients a snapshot of their visibility vs 9 competitors and has become an important benchmark for them in a language that marketers can relate to and enables them to tell a simple but effective story to their C-suite audience.
No problem Keith. I really like this metric for competitor analysis too; one of the best ways of estimating traffic.
Interesting. I don't think a lot of people take into account the value of each keyword's rank in the SERPs. I've read studies that showed the difference between the first rank and the second was a difference of 33% (first ranking) in comparison to 18% (second ranking). That's a significant difference from first and second place! I guess that's why your metrics on share of voice is rather sound.
Thanks for showing us how to do the math, Peter! It's a lot simpler than it seems at first.
Yeah, it's really important to understand the difference in CTRs of the different SERP positions. It becomes apparent why the top spots are so important... However, much of this is changing with the introduction of space-stealing elements such as extended Ads and so on. Will be interested in seeing how this changes.
Brilliant post. I love this topic. It came up when I attended Jessica Bowman's InHouse SEO Exchange in Santa Monica last May. The reason it came up was I asked the group how to deal with a CEO who was fixated on the fact that one of our competitors outranked us for a particular keyword. Sure, they outranked us, but they were #13 and we were #18 in Google. I was trying to get him to understand that, aside from simple vanity, there was no real competitive advantage they were gaining over us. After all it, neither of us were close to being on page 1.
The beauty of what you've done here is give SEOs, at least at a small scale, the ability to illustrate with data what to my CEO at the time was only my "opinion." According to your chart, the click-thru rates for spots #13 and #18 are identical, meaning organic market share for either position would be identical.
That's solid information and data that I can take to a stakeholder and explain in business terms they understand. CEOs "get" market share. When you can show them that they need to lift their eyes a bit and look down the road instead of staring at the hood-ornament, hopefully you can get them to understand that obsessing over a competitor that's 3 spots above you on page 2 of Google is not worthwhile. Creating a game plan to improve where you're at is a much better fixation.
Thanks Peter!
Thank you very much Dana!
Yes, it's important to understand the relative strength of different SERP positions. It's even better when you can arm yourself with the broad understanding a share of voice calculation will give you. If you're competing with a number of different competitors in a particular niche, if you can show share of voice for a chunk of important keywords and track this over time, you'll have a really good idea of how you're doing compared to them.
The way you calculate the Share of Voice is very interesting. Even though there are more ways to do that I can say that calculating the share of voice for each keyword it's more detailed, and as long as it's more detailed I think that it's also more effective.