Tag Archive | "Measuring"

Measuring the quality of popular keyword research tools

Contributor JR Oakes measures the quality of popular keyword research tools against data found in Google search results and performing page data from Google Search Console.



Please visit Search Engine Land for the full article.


Search Engine Land: News & Info About SEO, PPC, SEM, Search Engines & Search Marketing

Posted in IM NewsComments Off

Measuring SEO Performance After “Not Provided”

In recent years, the biggest change to the search landscape happened when Google chose to withhold keyword data from webmasters. At SEOBook, Aaron noticed and wrote about the change, as evermore keyword data disappeared.

The motivation to withold this data, according to Google, was privacy concerns:

SSL encryption on the web has been growing by leaps and bounds. As part of our commitment to provide a more secure online experience, today we announced that SSL Search on https://www.google.com will become the default experience for signed in users on google.com.

At first, Google suggested it would only affect a single-digit percentage of search referral data:

Google software engineer Matt Cutts, who’s been involved with the privacy changes, wouldn’t give an exact figure but told me he estimated even at full roll-out, this would still be in the single-digit percentages of all Google searchers on Google.com

…which didn’t turn out to be the case. It now affects almost all keyword referral data from Google.

Was it all about privacy? Another rocket over the SEO bows? Bit of both? Probably. In any case, the search landscape was irrevocably changed. Instead of being shown the keyword term the searcher had used to find a page, webmasters were given the less than helpful “not provided”. This change rocked SEO. The SEO world, up until that point, had been built on keywords. SEOs choose a keyword. They rank for the keyword. They track click-thrus against this keyword. This is how many SEOs proved their worth to clients.

These days, very little keyword data is available from Google. There certainly isn’t enough to keyword data to use as a primary form of measurement.

Rethinking Measurement

This change forced a rethink about measurement, and SEO in general. Whilst there is still some keyword data available from the likes of Webmaster Tools & the AdWords paid versus organic report, keyword-based SEO tracking approaches are unlikely to align with Google’s future plans. As we saw with the Hummingbird algorithm, Google is moving towards searcher-intent based search, as opposed to keyword-matched results.

Hummingbird should better focus on the meaning behind the words. It may better understand the actual location of your home, if you’ve shared that with Google. It might understand that “place” means you want a brick-and-mortar store. It might get that “iPhone 5s” is a particular type of electronic device carried by certain stores. Knowing all these meanings may help Google go beyond just finding pages with matching words

The search bar is still keyword based, but Google is also trying to figure out what user intent lays behind the keyword. To do this, they’re relying on context data. For example, they look at what previous searches has the user made, their location, they are breaking down the query itself, and so on, all of which can change the search results the user sees.

When SEO started, it was in an environment where the keyword the user typed into a search bar was exact matching that with a keyword that appears on a page. This is what relevance meant. SEO continued with this model, but it’s fast becoming redundant, because Google is increasingly relying on context in order to determine searcher intent & while filtering many results which were too aligned with the old strategy. Much SEO has shifted from keywords to wider digital marketing considerations, such as what the visitor does next, as a result.

We’ve Still Got Great Data

Okay, if SEO’s don’t have keywords, what can they use?

If we step back a bit, what we’re really trying to do with measurement is demonstrate value. Value of search vs other channels, and value of specific search campaigns. Did our search campaigns meet our marketing goals and thus provide value?

Do we have enough data to demonstrate value? Yes, we do. Here are a few ideas SEOs have devised to look at the organic search data they are getting, and they use it to demonstrate value.

1. Organic Search VS Other Activity

If our organic search tracking well when compared with other digital marketing channels, such as social or email? About the same? Falling?

In many ways, the withholding of keyword data can be a blessing, especially to those SEOs who have a few ranking-obsessed clients. A ranking, in itself is worthless, especially if it’s generating no traffic.

Instead, if we look at the total amount of organic traffic, and see that it is rising, then we shouldn’t really care too much about what keywords it is coming from. We can also track organic searches across device, such as desktop vs mobile, and get some insight into how best to optimize those channels for search as a whole, rather than by keyword. It’s important that the traffic came from organic search, rather than from other campaigns. It’s important that the visitors saw your site. And it’s important what that traffic does next.

2. Bounce Rate

If a visitor comes in, doesn’t like what is on offer, and clicks back, then that won’t help rankings. Google have been a little oblique on this point, saying they aren’t measuring bounce rate, but I suspect it’s a little more nuanced, in practice. If people are failing to engage, then anecdotal evidence suggests this does affect rankings.

Look at the behavioral metrics in GA; if your content has 50% of people spending less than 10 seconds, that may be a problem or that may be normal. The key is to look below that top graph and see if you have a bell curve or if the next largest segment is the 11-30 second crowd.

Either way, we must encourage visitor engagement. Even small improvements in terms of engagement can mean big changes in the bottom line. Getting visitors to a site was only ever the first step in a long chain. It’s what they do next that really makes or breaks a web business, unless the entire goal was that the visitor should only view the landing page. Few sites, these days, would get much return on non-engagement.

PPCers are naturally obsessed with this metric, because each click is costing them money, but when you think about it, it’s costing SEOs money, too. Clicks are getting harder and harder to get, and each click does have a cost associated with it i.e. the total cost of the SEO campaign divided by the number of clicks, so each click needs to be treated as a cost.

3. Landing Pages
We can still do landing page analysis. We can see the pages where visitors are entering the website. We can also see which pages are most popular, and we can tell from the topic of the page what type of keywords people are using to find it.

We could add more related keyword to these pages and see how they do, or create more pages on similar themes, using different keyword terms, and then monitor the response. Similarly, we can look at poorly performing pages and make the assumption these are not ranking against intended keywords, and mark these for improvement or deletion.

We can see how old pages vs new pages are performing in organic search. How quickly do new pages get traffic?

We’re still getting a lot of actionable data, and still not one keyword in sight.

4. Visitor And Customer Acquisition Value

We can still calculate the value to the business of an organic visitor.

We can also look at what step in the process are organic visitors converting. Early? Late? Why? Is there some content on the site that is leading them to convert better than other content? We can still determine if organic search provided a last click-conversion, or a conversion as the result of a mix of channels, where organic played a part. We can do all of this from aggregated organic search data, with no need to look at keywords.

5. Contrast With PPC

We can contrast Adwords data back against organic search. Trends we see in PPC might also be working in organic search.

For AdWords our life is made infinitesimally easier because by linking your AdWords account to your Analytics account rich AdWords data shows up automagically allowing you to have an end-to-end view of campaign performance.

Even PPC-ers are having to change their game around keywords:

The silver lining in all this? With voice an mobile search, you’ll likely catch those conversions that you hadn’t before. While you may think that you have everything figured out and that your campaigns are optimal, this matching will force you into deeper dives that hopefully uncover profitable PPC pockets.

6. Benchmark Against Everything

In the above section I highlighted comparing organic search to AdWords performance, but you can benchmark against almost any form of data.

Is 90% of your keyword data (not provided)? Then you can look at the 10% which is provided to estimate performance on the other 90% of the traffic. If you get 1,000 monthly keyword visits for [widgets], then as a rough rule of thumb you might get roughly 9,000 monthly visits for that same keyword shown as (not provided).

Has your search traffic gone up or down over the past few years? Are there seasonal patterns that drive user behavior? How important is the mobile shift in your market? What landing pages have performed the best over time and which have fallen hardest?

How is your site’s aggregate keyword ranking profile compared to top competitors? Even if you don’t have all the individual keyword referral data from search engines, seeing the aggregate footprints, and how they change over time, indicates who is doing better and who gaining exposure vs losing it.

Numerous competitive research tools like SEM Rush, SpyFu & SearchMetrics provide access to that type of data.

You can also go further with other competitive research tools which look beyond the search channel. Is most of your traffic driven from organic search? Do your competitors do more with other channels? A number of sites like Compete.com and Alexa have provided estimates for this sort of data. Another newer entrant into this market is SimilarWeb.

And, finally, rank checking still has some value. While rank tracking may seem futile in the age of search personalization and Hummingbird, it can still help you isolate performance issues during algorithm updates. There are a wide variety of options from browser plugins to desktop software to hosted solutions.

By now, I hope I’ve convinced you that specific keyword data isn’t necessary and, in some case, may have only served to distract some SEOs from seeing other valuable marketing metrics, such as what happens after the click and where do they go next.

So long as the organic search traffic is doing what we want it to, we know which pages it is coming in on, and can track what it does next, there is plenty of data there to keep us busy. Lack of keyword data is a pain, but in response, many SEOs are optimizing for a lot more than keywords, and focusing more on broader marketing concerns.

Further Reading & Sources:

Categories: 

SEO Book

Posted in IM NewsComments Off

Measuring Social Media

Measuring PPC and SEO is relatively straightforward. But how do we go about credibly measuring social media campaigns, and wider public relations and audience awareness campaigns?

As the hype level of social media starts to fall, then more questions are asked about return on investment. During the early days of anything, the hype of the new is enough to sustain an endeavor. People don’t want to miss out. If their competitors are doing it, that’s often seen as good enough reason to do it, too.

You may be familiar with this graph. It’s called the hype cycle and is typically used to demonstrate the maturity, adoption and social application of specific technologies:

Where would social media marketing be on this graph?

I think a reasonable guess, if we’re seeing more and more discussion about ROI, is somewhere on the “slope of enlightenment”. In this article, we’ll look at ways to measure social media performance by grounding it in the only criteria that truly matter – business fundamentals.

Public Relations

We’ve talked about the Cluetrain Manifesto and how the world changed when corporations could no longer control the message. If the message can no longer be controlled, then measuring the effectiveness of public relations becomes even more problematic.

PR used to be about crafting a message and placing it, and nurturing the relationships that allowed that to happen. With the advent of social media, that’s still true, but the scope has expanded exponentially – everyone can now repeat, run with, distort, reconfigure and reinvent the messages. Controlling the message was always difficult, but now it’s impossible.

On the plus side, it’s now much easier to measure and quantify the effectiveness of public relations activity due to the wealth of web data and tools to track what people are saying, to whom, and when.

The Same, Only Different

As much as things change, the more they stay the same. PR and social media is still about relationships. And getting relationships right pays off:

Today, I want to write about something I’d like to call the “Tim Ferriss Effect.” It’s not exclusive to Tim Ferriss, but he is I believe the marquee example of a major shift that has happened in the last 5 years within the world of book promotion. Here’s the basic idea: When trying to promote a book, the main place you want coverage is on a popular single-author blog or site related to your topic…..The post opened with Tim briefly explaining how he knew me, endorsing me as a person, and describing the book (with a link to my book.) It then went directly into my guest post– there was not even an explicit call to action to buy my book or even any positive statements about my book. An hour later, (I was #45 on Amazon’s best seller list

Public relations is more than about selling, of course. It’s also about managing reputation. It’s about getting audiences to maintain a certain point of view. Social media provides the opportunity to talk to customers and the public directly by using technology to dis-intermediate the traditional gatekeepers.

Can We Really Measure PR & Social Media Performance?

How do you measure the value of a relationship?

Difficult.

How can you really tell if people feel good enough about your product or service to buy it, and that “feeling good” was the direct result of editorial placement by a well-connected public relations professional?

Debatable, certainly.

Can you imagine another marketing discipline that used dozens of methods for measuring results? Take search engine marketing for example. The standards are pretty cut and dry: visitors, page views, time on site, cost per click, etc. For email marketing, we have delivery, open rates, click thru, unsubscribes, opt-ins, etc”

In previous articles, we’ve looked at how data-driven marketing can save time and be more effective. The same is true of social media, but given it’s not an exact science, it’s a question of finding an appropriate framework.

There are a lot of people asking questions about social media’s worth.

No Industry Standard

Does sending out weekly press releases result in more income? How about tweeting 20 times a day? How much are 5,000 followers on Facebook worth? Without a framework to measure performance, there’s no way of knowing.

Furthermore, there’s no agreed industry standard.

In direct marketing channels, such as SEO and PPC, measurement is fairly straightforward. We count cost per click, number of visitors, conversion rate, time on site, and so on. But how do we measure public relations? How do we measure influence and awareness?

PR firms have often developed their own in-house terms of measurement. The problem is that without industry standards, success criteria can become arbitrary and often used simply to show the agency in a good light and thus validate their fees.

Some agencies use publicity results, such as the number of mentions in the press, or the type of mention i.e. prestigious placement. Some use advertising value equivalent i.e. is what editorial coverage would cost if it were buying advertising space. Some use public opinion measures, such as polls, focus groups and surveys, whilst others compare mentions and placement vs competitors i.e. who has more or better mentions, wins. Most use a combination, depending on the nature of the campaign.

Most business people would agree that measurement is a good thing. If we’re spending money, we need to know what we’re getting for that money. If we provide social media services to clients, we need to demonstrate what we’re doing works, so they’ll devote more budget to it in future. If the competition is using this channel, then we need to know if we’re using it better, or worse, than we are.

Perhaps the most significant reason why we measure is to know if we’ve met a desired outcome. To do that we must ignore gut feelings and focus on whether an outcome was achieved.

Why wouldn’t we measure?

Some people don’t like the accountability. Some feel more comfortable with an intuitive approach. It can be difficult for some to accept that their pet theories have little substance when put to the test. It seems like more work. It seems like more expense. It’s just too hard. When it comes to social media, some question whether it can be done much at all

For proof, look no further than The Atlantic, which shook the social media realm recently with its expose of “dark social” – the idea that the channels we fret over measuring like Facebook and Twitter represent only a small fraction of the social activity that’s really going on. The article shares evidence that reveals that the vast majority of sharing is still done through channels like email and IM that are nearly impossible to measure (and thus, dark).

And it’s not like a lot of organizations are falling over themselves to get measurement done:

According to a Hypatia Research report, “Benchmarking Social Community Platform Investments & ROI,” only 40% of companies measure social media performance on a quarterly or annual basis, while almost 13% or the organizations surveyed do not measure ROI from social media at all, and another 18% said they do so only on an ad hoc basis. (Hypatia didn’t specify what response the other 29% gave.)

If we agree that measurement is a good thing and can lead to greater efficiency and better decision making, then the fact your competition may not be measuring well, or at all, then this presents great opportunity. We should strive to measure social media ROI, as providers or consumers, or it becomes difficult to justify spend. The argument that we can’t measure because we don’t know all the effects of our actions isn’t a reason not to measure what we can.

Marketing has never been an exact science.

What Should We Measure?

Measurement should be linked back to business objectives.

In “Measure What Matters”, Katie Delahaye Paine outlines seven steps to social media measurement. I liked these seven steps, because they aren’t exclusive to social media. They’re the basis for measuring any business strategy and similar measures have been used in marketing for a long time.

It’s all about proving something works, and then using the results to enhance future performance. The book is a great source for those interested in reading further on this topic, which I’ll outline here.

1. What Are Your Objectives?

Any marketing objective should serve a business objective. For example, “increase sales by X by October 31st”.

Having specific, business driven objectives gets rid of conjecture and focuses campaigns. Someone could claim that spending 30 days tweeting a new message a day is a great thing to do, but if, at the end of it, a business objective wasn’t met, then what was the point?

Let’s say an objective is “increase sales of shoes compared to last December’s figures”. What might the social strategy look like? It might consist of time-limited offers, as opposed to more general awareness messages. What if the objective was to “get 5,000 New Yorkers to mention the brand before Christmas”? This would lend itself to viral campaigns, targeted locally. Linking the campaign to specific business objectives will likely change the approach.

If you have multiple objectives, you can always split them up into different campaigns so you can measure the effectiveness of each separately. Objectives typically fall into sales, positioning, or education categories.

2. Who Is The Audience?

Who are you talking to? And how will you know if you’ve reached them? Once you have reached them, what is it you want them to do? How will this help your business?

Your target audience is likely varied. Different audiences could be industry people, customers, supplier organizations, media outlets, and so on. Whilst the message may be seen by all audiences, you should be clear about which messages are intended for who, and what you want them to do next. The messages will be different for each group as each group likely picks up on different things.

Attach a value to each group. Is a media organization picking up on a message more valuable than a non-customer doing so? Again, this should be anchored to a business requirement. “We need media outlets following us so they may run more of our stories in future. Our research shows more stories has led to increased sales volume in the past”. Then a measure might be to count the number of media industry followers, and to monitor the number of stories they produce.

3. Know Your Costs

What does it cost you to run social media campaigns? How much time will it take? How does this compare to other types of campaigns? What is your opportunity cost? How much does it cost to measure the campaign?

As Delahaye Paine puts it, it’s the “I” in ROI.

4. Benchmark

Testing is comparative, so have something to compare against.

You can compare yourself against competitors, and/or your own past performance. You can compare social media campaigns against other marketing campaigns. What do those campaigns usually achieve? Do social media campaigns work better, or worse, in terms of achieving business goals?

In terms of ROI, what’s a social media “page view” worth? You could compare this against the cost of a click in PPC.

5. Define KPIs

Once you’ve determined objectives, defined the audience, and established benchmarks, you should establish criteria for success.

For example, the objective might be to increase media industry followers. The audience is the media industry and the benchmark is the current number of media industry followers. The KPI would be the number of new media industry followers signed up, as measured by classifying followers into subgroups and conducting a headcount.

Measuring the KPI will differ depending on objective, of course. If you’re measuring the number of mentions in the press vs your competitor, that’s pretty easy to quantify.

“Raising awareness” is somewhat more difficult, however once you have a measurement system in place, you can start to break down the concept of “awareness” into measurable components. Awareness of what? By whom? What constitutes awareness? How to people signal they’re aware of you? And so on.

6. Data Collection Tools

How will you collect measurement data?

  • Content analysis of social or traditional media
  • Primary research via online, mail or phone survey
  • Web analytics

There are an overwhelming number of tools available, and outside the scope of this article. No tool can measure “reputation” or “awareness” or “credibility” by itself, but can produce usable data if we break those areas down into suitable metrics. For example, “awareness” could be quantified by “page views + a survey of a statistically valid sample”.

Half the battle is asking the right questions.

7. Take Action

A measurement process is about iteration. You do something, get the results back, act on them and make changes, and arrive at a new status quo. You then do something starting from that new point, and so on. It’s an ongoing process of optimization.

Were objectives met? What conclusions can you draw?

Those seven steps will be familiar to anyone who has measured marketing campaigns and business performance. They’re grounded in the fundamentals. Without relating social media metrics back to the underlying fundamentals, we can never be sure if what we’re doing is making or a difference, or worthwhile. Is 5,000 Twitter followers a good thing?

It depends.

What business problem does it address?

Did You Make A Return?

You invested time and money. Did you get a return?

If you’ve linked your social media campaigns back to business objectives you should have a much clearer idea. Your return will depend on the nature of your business, of course, but it could be quantified in terms of sales, cost savings, avoiding costs or building an audience.

In terms of SEO, we’ve long advocated building brand. Having people conduct brand searches is a form of insurance against Google demoting your site. If you have brand search volume, and Google don’t return you for brand searches, then Google looks deficient.

So, one goal of social media that gels with SEO is to increase brand awareness. You establish a benchmark of branded searches based on current activity. You run your social media campaigns, and then see if branded searches increase.

Granted, this is a fuzzy measure, especially if you have other awareness campaigns running, as you can’t be certain cause and effect. However, it’s a good start. You could give it a bit more depth by integrating a short poll for visitors i.e. “did you hear about us on Twitter/Facebook/Other?”.

Mechanics Of Measurement

Measuring social media isn’t that difficult. In fact, we could just as easily use search metrics in many cases. What is the cost per view? What is the cost per click? Did the click from a social media campaign convert to desired action? What was your business objective for the social media campaign? To get more leads? If so, then count the leads. How much did each lead cost to acquire? How does that cost compare to other channels, like PPC? What is the cost of customer acquisition via social media?

In this way, we could split social media out into the customer service side and marketing side. Engaging with your customers on Facebook may not be all that measurable in terms of direct marketing effects, it’s more of a customer service function. As such, budget for the soft side of social media need not come out of marketing budgets, but customer service budgets. This could still be measured, or course, by running customer satisfaction surveys.

Is Social Media Marketing Public Relations?

Look around the web for definitions of the differences between PR and social media, and you’ll find a lot of vague definitions.

Social media is a tool used often used for the purpose of public relations. The purpose is to create awareness and nurture and guide relationships.

Public relations is sometimes viewed it as a bit of a scam. It’s an area that sucks money, yet can often struggle to prove its worth, often relying on fuzzy, feel-good proclamations of success and vague metrics. It doesn’t help that clients can have unrealistic expectations of PR, and that some PR firms are only too happy to promise the moon:

PR is nothing like the dark, scary world that people make it out to be—but it is a new one for most. And knowing the ropes ahead of time can save you from setting impossibly high expectations or getting overpromised and oversold by the firm you hire. I’ve seen more than my fair share of clients bringing in a PR firm with the hopes that it’ll save their company or propel a small, just-launched start-up into an insta-Facebook. And unfortunately, I’ve also seen PR firms make these types of promises. Guess what? They’re never kept

Internet marketing, in general, has a credibility problem when it doesn’t link activity back to business objectives.

Part of that perception, in relation to social media, comes from the fact public relations is difficult to control:

The main conduit to mass publics, particularly with a consumer issue such as rail travel or policing, are the mainstream media. Unlike advertising, which has total control of its message, PR cannot convey information without the influence of opinion, much of it editorial. How does the consumer know what is fact, and what has influenced the presentation of that fact?

But lack of control of the message, as the Cluetrain Manifesto points out, is the nature of the environment in which we exist. Our only choice, if we are to prosper in the digital environment, is to embrace the chaos.

Shouldn’t PR just happen? If you’re good, people just know? Well, even Google, that well known, engineering-driven advertising company has PR deeply embedded from almost day one:

David Krane was there from day one as Google’s first public relations official. He’s had a hand in almost every single public launch of a Google product since the debut of Google.com in 1999.

Good PR is nurtured. It’s a process. The way to find out if it’s good PR or ineffective PR is to measure it. PR isn’t a scam, anymore so than any other marketing activity is a scam. We can find out if it’s worthwhile only by tracking and measuring and linking that measurement back to a business case. Scams lack transparency.

The way to get transparency is to measure and quantify.

Categories: 

SEO Book

Posted in IM NewsComments Off


Advert