Webinar Wrap Up: Tracking the Trends in Digital Marketing for 2020 - A SEO Perspective

Changes in Matching in 2019

One of the biggest changes in search engine marketing in 2019 was to match types.  Both phrase and modified broad match underwent some changes.

If you've been using search engine marketing but haven't kept up with the changes, phrase match once always meant having to have the same words in the same order in a search query.  Google made two critical changes in 2019:

• Terms could suddenly be in different orders, and
• Google could substitute words with what they deemed as the same meaning as the keyword into the query.

This has led some companies to have some issues with match types.  For instance, a foreign travel company might have found that Google suddenly said “Kenya Safari” is the same thing as “Nairobi safari.”  Nairobi is indeed the capital of Kenya, but an exotic travel company knows that, from an average order value standpoint, they're vastly different: a safari throughout Kenya can involve many more experiences than Nairobi (and may not even begin or end in Nairobi).

Here’s another example: a search for “trademark registration cost” is usually from someone who is trying to find out the cost to have a third party register trademarks for them.  When someone searches “cost to file a trademark,” it's very different; it’s usually a person who wants to do the filing themselves.

We've seen some language substitutions.  We've seen some word substitutions that are completely different in intent: trademark logo versus trademark symbol, for example.  If you've been using a lot of phrase match, it can be beneficial to go back in your search term data and take a look at your close variant phrase matches.  Close variants are words that are not in your account that Google's matching. Check those words to see if you're having issues because you're mostly going to fix these with negative keywords.  Not everyone has these problems, but this problem type is one to look for because it was a big trend last year.

Second, the modified broad match changed.  Modified broad has always been a great match type to use: you didn't have to figure out every single word combination and every single keyword to use, but when you used the plus symbol, Google would never go crazy on their matching.  Last year, they changed the rules to allow similar word substitutions, and so you might see things like Google treating the word “trademark” as the same as “logo design.” However, logo design and trademarks are not even close in meaning in some contexts.

Another issue is matching to brand names.  Google rolled out changes to this last summer to fall in English, it rolled out to Germany in February-March this year and to the Netherlands in February to March this year (as well as some other countries).  Based on the languages you work with, you may not see these changes: they may still be rolling out to your accounts based upon language.

Ad Group and Keyword Side Effects 

If you had a good account organization and suddenly Google's saying, “yes,  you've got this keyword and ad group here, but we also think your groups 2, 3, & 4 have a good match to the same keyword,” suddenly you have duplicate queries happening.  In duplicate queries, the same query shows up from multiple ad groups. For example, in figure 1, there are groups called “Register trademark,” “Trademark logo,” and “Trademarking.”  In this case, “Trademarking” is meant to catch queries that don’t exactly match the keywords in the other groups.

Figure 1.  Ad Group and Keyword Query Duplication

The “Register trademark” group has a keyword “registered trademark,” and they had 

Eleven impressions in a month for that search term under an exact match type. But the same search term showed up in over a thousand times in the “Trademarking” group, matching the keyword “trademarking” in a broad match.

This is an important issue.  Often, when you have an exact phrase match in an ad group, you have an ad in a landing page highly reflective of those keywords, optimized for reasonable conversion rates.  If you have a backup ad group to capture the rest of the less-exact query volume, those ads and landing pages are often not as precise. If that same search term gets directed to the backup group rather than the precisely targeted group, you will get lower conversion rates.

You don't set bids by search terms; you set bids by keywords or by ad groups.  Testing your bid systems gets badly messed up if the same search term is in several or many different ad groups. You want to discover where these exist so you can fix them.

By comparing earlier years to the current year, you can see if this kind of problem exists.  If you go into your keywords, download your data, and create a pivot table, you can look for changes to your CPAs year over year or month over month.  If your CPAs increase sharply for specific keywords, there's usually a problem. You can repeat this exact process for search query data instead of keywords.  Pivot tables are a powerful tool for this analysis. 

From Average Position to Click Share to Conversions

In 2019, Google retired the average position metric and then introduced click share to go with impression share.  While everyone has been forced to learn impressions share in the last few months because the average position is gone, click share is still an underused metric.

Click share is the percentage of clicks you received compared to what was possible.  You want to start graphing click share over time to see how it trends against impression numbers.  You can then begin to see how well click share tracks impression share metrics. Does it track top impression share, for example?  Does it mirror lost impressions? This gives you clues on what actions to take to increase click share.

You also want to track conversions against these metrics.  Conversions are, after all, the goal of your search engine marketing.  Another subtle change Google made that has huge implications is campaign level conversion tracking.

Some ecommerce companies have some campaigns for sales and others for engagement only.  For example, let's say you're a business to business company and you've got one campaign whose goal is a white paper downloads, another campaign whose goal is product demos, and a third one that's just direct sales.  On your website, one landing page converts to white paper downloads, another one converts to the demo, and another produces actual sales revenue. You wouldn't want all three of those to be used in an automated bid system based on revenue generation: you wouldn't want to count a white paper download or a demo as a conversion in a revenue-based campaign.

With the new settings, you can say for the white paper campaign, and we only want to use the white paper conversion tracking, for the demo campaign demos, and the revenue generation campaign actual revenues.  You can track your campaigns differently by their different purposes.

Another campaign-level conversion tracking change is seasonal bid adjustments.  If you're using robust bidding, a problem arises when you know a seasonal event like Black Friday will increase sales and will impact your bidding system.  The bidding algorithms are still not very good at predicting significant upswings and conversion rates and how those affect bids.

To compensate, you can specify seasonal adjustments.  If you know in a given time frame your conversions will spike, you can tell the machine, “don't look only at our past data in determining our future bids: look at our past data plus an adjustment for this time frame in determining bids.”

Other new campaign-related settings include specifying a maximum conversion value and setting a target impression share.  In reality, target impression share bidding is the same as the old bid to position, but since the average position was retired, target impression share is the replacement metric.

Other Minor Changes

There were several other minor changes in Google in 2019.

• An improved keyword planner.
• A new performance planner.  This tool combines sets of proposed keywords with proposed project spend to project what results to expect. 
• Retirement of accelerated delivery (the option to spend your budget as fast as possible).
• Various ad extensions remain in beta, including the lead form, which has now been in beta for 14-15 months.
• New optimization scores that can impact you if you're in the Google partner program.

Predicted Trends in 2020 and How They’re Developing

One of the top predictions in 2020 was that business-to-business would finally get some help.  It's always been a struggle, outside of LinkedIn and some Facebook audiences, to do well.  B2B in Google is still in beta - a very open beta, so if you're in B2B and you work with a Google rep, you can ask for this beta if you're not in it.

• One aspect of this B2B targeting is company size, where you want to target companies in a particular size range.  The number of companies in the targeting range is still fairly small. However, if you have different messages that you target to small companies versus the ones you target to large enterprises, you can use audience layers and now use the same keywords in different ads to these different sizes of companies.
• Another aspect is industry targeting, where you can target (for example) companies in construction or finance or hospitality.  These additional audience options are more useful for B2B than consumer-based accounts.

There are a few audience targeting options that came out of privacy laws, like GDPR in Europe or the California Privacy Act.  Everyone worried about how these laws affect digital marketing from a legal and technical standpoint. The good news is that the issue has been fixed in large parts, and now the big worry is how cookieless tracking will happen. Firefox, Safari, Chrome, and other browsers will start blocking third-party cookies.  Google, Microsoft, Facebook, Amazon, and Apple all have a pretty good grasp of how this will work, so you don't have to worry about the cookieless tracking long as it satisfies the privacy acts. However, as the legal issues and the technical issues are handled, we may see a return to looking for ways to segment audiences legally based upon all the current privacy rules.

Google has realized they probably went too far with phrase match word substitution, and they’re going to pull that back a little.  Exact match will probably get a little less relevant at times because it will get a bit more aggressive in word substitution. It doesn't use the phrase match rules for word substitution right now, but it will probably change.  The modified broad match should get more relevance again: this is where Google realized they went too far, and there's been some conversion rate pushback from advertisers.

The automation versus human debate is going to keep growing.  This debate has been happening for four to five years now, and it’s something you should watch because it will impact your career.  Companies are starting to figure that if they just use people to address search marketing issues, they miss scale too much - but if they only use computers, they miss out on human interaction and creativity.  Over time, every study has shown that when humans and computers intersect, marketing is much more efficient so that we will see more blending of the two.

Microsoft is making big improvements in search marketing tools.  If you're still on the fence about Bing ads or Microsoft ads, you should be using them. More significant is the prediction that LinkedIn targeting should leave beta this year. This means you will be able to pull LinkedIn targeting data directly into your Microsoft ads accounts.  We are going to see different ad extensions roll out from Microsoft this year: we should see more syndication relationships from them, which means more impressions for your ads. The biggest complaint about using Microsoft ads is just lack of inventory, and they're hard at working on syndication relationships to increase that impression volume.  We should see some announcements this year in their effort to achieve parity with Google.

On the career front, the good news is that according to McKinley Marketing, sixty-one percent of companies plan to add digital marketers to their team this year. The biggest problem these companies face is too few quality candidates: there are not enough qualified people for their open jobs.  Right now, due to the COVID-19 crisis, we do not see the expansion that was expected, but once the crisis is over and things get back to a more normal state, there should be a huge help Wanted sign as companies rush to spend budgets left over. And as some companies receive bailout funds, many will put this money into marketing.

The COVID-19 crisis has made e-commerce explode.  Ecommerce spending in recent weeks was up to eighty percent compared to the same period last year, and ecommerce digital marketing spending has increased accordingly due to people abandoning retail stores and ordering online.  This is a good time to focus on the skills you need so that when things are back to a non-isolation environment, you are ready to take the jobs that are going to open up.

The Trends We All Missed

As much as people like to predict trends, there are always trends that we miss.  Number one this year is the COVID-19 virus. Nobody no one saw this coming, and it’s hit different industries in different ways.  For example, if you're involved in a travel-based business, your marketing spend probably dropped to near zero. By contrast, if you're involved in ecommerce, you're probably doing more volume than you've ever seen before.  If your company specializes in remote work, like Zoom or GoToMeeting, this is a great time to be marketing them. If you're selling staple goods, it’s a good time to market them; if you're in luxury goods, you're seeing a decline.  Regardless of what industry you work in, you probably have a very different time at work then you did a month ago.

Ecommerce growth has increased significantly more than any projections.  The brilliant people who crunch numbers to make e-commerce projections to plan marketing budgets have all severely undervalued ecommerce this year. 

With everyone working home right now and companies not spending much on travel, the big question is. how does that affect the work environment once the virus passes?  Will there be more virtual companies and more flexible work hours? Initially, right after the self-isolation ends, we will see everyone sort of going back to work because no one's prepared yet to say they really should be more flexible on things.  But over time, if companies decide they don't need a big building with 50,000 people in it anymore and can shift to 20,000 people in the office and 30,000 remote workers, what does the Home Office change look like? We're going to see some big changes to how remote work happens in the future with more of a GoToMeeting, Zoom, WebEx conference world versus all the in-person events of past years.

This has some weird implications for search marketing.  For instance, big companies have many more remote workers using home networks.  Most companies of that size have filters in place in their analytics to filter out their corporate IP addresses and their VPNs, so they don't affect their analytics data.  What they don't have is a list of every single employee’s home IP address, so now their workers who used to be in the office and were excluded from analytics are at home using the company website as usual, and suddenly they are now counted in the analytics.  These companies may wonder why their conversion rates are down so much, but their sales are the same. A trend like that can result when they’ve polluted their analytics due to their remote workers using their site, adding more total sessions, which shouldn't be counted in the conversion rate.  Hidden side effects like that go along with working from home, so we need to think beyond just how to be productive in a non-office environment. We need to make sure that as we get everything done through working from home, we compensate for things that affect our data in new, interesting, and unexpected ways. 

You can also watch the webinar video to know more about these trends in digital marketing from 2019 through 2020.

Key Takeaways

Brad identified key takeaways from the webinar:

Changes in match types.  Every match type except broad, has been changed in the past two and a half years. so looking through your search query data for duplicate keywords, duplicate queries, and poor intent matching will help you find keyword matches that should not exist.  You can get rid of those using your negative keywords wisely to make sure you're only showing ads on keywords and on search queries that have high intent.

Impression and click share. Monitoring how these trends over time will help you see how your data changes and how it's correlated.

Campaign level conversions. The ability to switch what conversions are counted by the campaign, especially when you use automated bidding, is an important change.

Google's optimization score.  Be careful: some of their suggestions are useful, but others are not and may not help your account.

More B2B targeting options.   LinkedIn a Microsoft will grow together: we should see the LinkedIn targeting leaves beta.

Subtle shifts of how match types work.  Not all of these will be announced by Google.  If they refine the match type algorithm to show more related queries and get rid of unrelated ones, that should increase people's conversion rates:  they're not likely to announce that because basically. they're tweaking machine learning. You may see the results in your data and not see actual announcements for the changes that caused them.

Your career in digital marketing. Going forward, one of the biggest issues companies will have is hiring qualified candidates once we're past the COVID-19 crisis.  They will be looking for candidates certified in skills like PPC, email marketing, mobile marketing, or a complete suite of digital marketing skills.  Making sure you've got the certifications that demonstrate you're a qualified candidate will help your career going forward once we get back to (relatively) normal.