SEM is undoubtedly one of the most effective digital marketing channels that exist. When the right elements come together – a great product with good unit economics, relatively few competitors, and high search volumes – paid search scales very nicely.
But startups rarely have all these elements working in their favor at once, and they usually face a set of unique challenges when trying to build out SEM as a scalable channel. As a digital marketing agency that has worked with more than 100 startups of all shapes and sizes on SEM, we have seen and navigated these issues and want to share some insights about how to overcome them.
Startups have a set of unique challenges above and beyond the typical challenges faced by companies advertising on paid search channels like Google AdWords and Bing Ads.
The first big challenge is that many startups have unknown or unproven customer LTVs (lifetime values). Knowing your LTVs is the most important first step when considering doing any kind of paid advertising. If you don’t know how much money you’re making from a customer over their lifetime, you won’t know how much you can afford to spend to acquire them.
But for startups with relatively little operating history and low customer volumes, it becomes very difficult to accurately project LTV. A lack of LTV data is a bit of a chicken and an egg problem, since LTV can only be established once you have acquired customers. You have to expend some combination of time, effort, and money to acquire customers from any channel.
This is very important to understand – SEM can be used to help validate or establish LTVs, but if this is your goal then make sure you are clear on what you are trying to achieve. Sometimes a startup thinks they are trying to see if SEM is scalable for their business, but they realize they don’t have the right LTV data to assess the efficiency of the channel in the first place. It’s an important distinction that will change the approach and the resources needed to manage your SEM accounts.
One way to get around this catch-22 is to get industry comparisons by talking to VCs or people at similar companies who can augment any customer data you have. While customer LTVs for other companies will likely differ from yours, it’s a relatively good measure that can at minimum help you establish a benchmark for developing a target CPA.
How much you can afford to spend to acquire a customer (called a target CPA) is directly related to how much scale you can achieve from SEM. The higher your target CPA, the more scale you can achieve.The reason for this is diminishing returns. As you start increasing your SEM spend, CPAs also increase as you saturate the low-hanging, high-intent keywords that typically convert well and you reach for new keyword opportunities. Eventually your actual CPA will start bumping up against your CPA target and you will be forced to stop scaling spend until either you determine you can afford to pay a higher CPA or you are able to achieve lower CPAs with optimization.
Another challenge for startups using paid search is heavy competition from established companies. Because these incumbents know their LTVs and also typically have much larger budgets to work with, they can afford to be more aggressive with their bids on SEM and this can squeeze out startups trying to break in. In addition, sometimes these companies ensure they show up in a certain position – regardless of how inefficient it may be. This can make it very difficult for scrappy, unit-economics-driven startups to compete. This can be exacerbated for high-funnel keywords that are already very expensive and have relatively low conversions rates.
For an eCommerce startup selling subscription clothing boxes, a generic high-funnel keyword like “mens jackets” is unlikely to back out to a target CPA that’s based on their average customer LTV. However, when searching for this term you’ll notice Ralph Lauren, Nordstrom, and Michael Kors bidding aggressively:
Even for these companies, it’s unlikely that the CPA from these keywords backs out for them directly, but they continue to spend millions of dollars per year to show up in the top positions for these keywords. Such companies may be bidding on the belief that the brand exposure they receive from showing up for generic keywords makes up for CPA inefficiency, or they may be bidding on these keywords purely due to a fear of missing out since their competitors are bidding on them. Regardless, this is not a battle that a startup is going to win unless their LTV, and thus target CPA, allows them to compete.
The best way to get around this is to bid on a variety of long-tail keywords, which will have lower competition and likely higher intent, but will also be cheaper. The more extensive your list of long-tail generic keywords, the more traffic you can drive without relying on competitive, generic keywords. Google’s Keyword Planner and third-party tools like SpyFu can help you find long-tail keywords with low competition that are relevant to your business.
Another common obstacle for startups looking to leverage paid search is that they’re often operating in nascent or emerging markets, leading to low search interest. This can limit the amount of relevant search volume available, which is a critical aspect to scaling paid search.
People have to know something exists before they can search for it, or at least something closely related to it. If a startup is selling a whole new class of product or service, this poses a challenge. For example, in the early days of Uber and Lyft, no one knew what on-demand ride sharing was, so no one was searching for it. Only as the market matured did search interest start to skyrocket, as shown below:
How can you access search volume at scale when your industry is so new that people don’t even know it exists yet? This can be a daunting obstacle, and it takes some creativity to get past it.
Targeting competitor keywords and tangential products is one way to do this. For example, in the early days Uber could have bid on competitor keywords like “avis car rental” and “yellow cab taxi”, or tangential products like “quick transportation” and “call a taxi” to build awareness of its product and drive conversions via SEM.
However, the relevancy of a keyword to the product at hand will determine the conversion rate and so again it comes down to what CPA you can afford to pay for these keywords. Waiting for your market to develop is sometimes necessary in order to truly achieve scale from SEM while hitting your CPA target.
We’ve addressed the challenges in making paid search a profitable and scalable channel for startups, but what about the light at the end of the tunnel for those companies that are able to overcome those challenges?
The most obvious and most important advantage of paid search is the value of intent. Users are telling you exactly what they’re looking for when they type in a search term, which usually indicate higher levels of purchase intent than indirect signals like audience targeting that are available for other paid acquisition channels. As a result, no paid advertising channels come close to paid search conversion rates.
Intent-based search ads allow you to capture the low hanging fruit. Someone searching for “ride sharing app” is already generally aware of your product and is probably deep enough in the marketing funnel to download the app and use it. Search ads let you maximize demand capture.
But intent is also valuable for capturing higher-funnel interest. Someone searching for “quick transportation” is sending a direct signal that they’re looking for a fast way to get around, even if they’re not aware of the existence of ride sharing. This goes to show that you can find conversion volume from keywords that are not directly related to your product or service offering, assuming the unit-economics work out when comparing LTV to CPA.
Brand keyword coverage plays an important role in all SEM programs, but it’s especially critical for startups who should be capturing everyone searching for them. All the effort and resources expended driving awareness of a startup can be wasted if a potential customer searches for your brand on Google and gets poached by a competitor’s search ad. Branded search ads can help you protect your investment in other channels.
Given how valuable these earned branded search queries are, it should come as no surprise that bidding on competitors’ keywords can be a very effective way to reach potential customers at the bottom of the funnel who are ready to make a decision. This can be especially beneficial for startups who are in markets with well-known incumbents (i.e. lots of customers searching for their branded keywords).
While established companies may have larger budgets and more brand name recognition working in their favor, startups also have some things working in their favor.
For one, startups are able to move very fast and learn quickly due to their small size and lack of red tape, allowing for high-tempo testing and optimization, which is for scaling a paid search program. For another, startups can prioritize growth over profitability in pursuit of capturing a market, allowing them to better compete with established companies in ad auctions.
Finally, startups tend to have new and refined product offerings compared to their older competitors, creating the potential for higher conversion rates and customer LTVs. This can make it easier to compete with established competitors.
A Word to the Wise – Don’t Ignore Paid Search
Startups eventually have to acquire new customers efficiently and at scale in order to become mainstay businesses. In the early days that’s no easy task, as even defining efficiency goals based on LTV benchmarks can be arduous. Succeeding with SEM as a startup requires being nimble with testing and learning, and it sometimes necessitates sacrificing short-term efficiency in favor of long-term growth.
There will always be established competitors with much larger budgets trying to box out startups trying to break into their market, but there’s a huge opportunity for the startups that are able to successfully overcome the challenges associated with SEM. People expressing their intent through a search query is arguably the most powerful marketing signal that exists. And tapping into that signal using paid search engine marketing can be the source of a highly scalable customer acquisition channel.
Here at BigFlare.com we’ve reviewed over 100 Google Shopping campaigns over the years. In that time, we’ve seen pretty much every money-wasting mistake that it’s possible to make in Shopping. Some of these are rare, one-time-only type errors. But some of these are surprisingly common. And, in fact, there are 3 mistakes in particular that we see being made in almost every account we review.
With this in mind, there is both good and bad news for Shopping advertisers. The bad news is that these mistakes are common and probably costing you money. The good news is, these mistakes are quick to fix.
A few clicks here, some minor adjustments there and presto! You should have a more profitable Shopping campaign in less than half a day’s work. So, let’s dig into each of the mistakes and the solutions for them.
Mistake 1: Not Bidding Per Product / Item ID
When we take over Shopping campaigns that have previously been run in-house, this is by far the most common type of error we see:
What we’re seeing here is one of the main ad groups within this client’s Shopping campaign, before we came along and optimised it. Within the ad group we have the product group. Bids are controlled at the product group level and you can split out your feed however you want. You could have all products in your feed in one product group, or you could have one product group per category, one product group per item, etc.
The mistake we often see, the one you’re seeing in the above image, is having the product groups split out into any level higher than the single product / item ID / SKU level. In the above example we see that all products are in one big product group for the whole account. Another common error we see is having one product group per category.
Whenever you are lumping lots of different products into the same product group you reduce your ability to bid optimally per product. CPC bids are controlled at the product group level. This means that if you have 1 product group with 84 products in it, you only have 1 CPC bid for those 84 products.
Your products will all tend to have different prices, conversion rates, click-through rates, and margins. Each one will thus have its own unique ideal CPC bid, and this ideal bid will change over time as the competitive landscape shifts and search volumes naturally fluctuate.
So, don’t limit yourself to only 1 CPC bid to cover multiple products. Split out your product groups as finely as possible: right down to the individual item level. This way, you can give each product the individually customised CPC bid that it needs and deserves.
The criminal part of all this is that the one-product-per-product-group structure is REALLY easy to set up. You just need to subdivide your product group:
And then select Item ID as the option to subdivide by:
This would give you a very simple structure of having 1 ad group with all your products in it, and each product having its own product group. There are many other ways of structuring your ad groups and this is just one of them, the simplest one. What we’ll show next is the next level of ad group / product organisation, the SPAG.
Taking This 1 Step Further: The SPAG
SPAG stands for Single Product Ad Groups. This is taking the one-product-per-ad-group structure even further and splitting it out so we have one product in each ad group. If you have more than a few products in your feed, this is going to take a while to do, so I’d recommend using a tool such as PPCsamurai.com or Optmyzr.com to help you do this quickly.
The main benefit of splitting out your Shopping campaign into SPAGs is a highly increased ability to “funnel” search terms using negative keywords.
Let’s say, for example, you have two products, product A and product B, that would match the query “blue trainers”. Product A is low margin and less profitable, product B is a bestseller with high margins and conversion rates. Of course, if someone searches “blue trainers” you would prefer them to see Product B’s ad, not product A.
Without a SPAG structure you could try to do this by setting bids appropriately, but it’s never going to be foolproof. With a SPAG structure you can now create a negative keyword in A’s ad group for “blue trainers” and all the traffic for that keyword will now correctly be funneled to product B. If you don’t split out into SPAGs, you lose the ability to do this, as negative keywords are controlled at the ad group level, not the product group.
Mistake 2: Not Splitting out Brand vs Non-Brand
You’re already splitting out your brand traffic from your non-brand traffic in your Search campaigns, right? Please tell me you are! Brand traffic is an entirely different beast to non-brand traffic. It consists of people who already know your brand and want to come directly to your site and (hopefully) buy something. Non-brand traffic will contain a much higher proportion of new vs returning visitors, and thus it needs to be treated differently to Brand, different targets, different bids, etc.
So that’s why every search marketer worth their salt is already splitting out brand vs non-brand traffic in Search. Why isn’t everyone also doing it in Shopping? I think the answer is that, because Shopping does not have keyword targeting, most Shopping advertisers do not even realise that splitting out brand vs non-brand traffic is even an option. But it is!
Using a clever campaign priority strategy combined with the correct negative keywords, you can indeed split out Shopping into brand vs non-brand searches. Here’s how:
1. Create Your Brand Campaign
This is going to be an exact duplicate of your normal “Generics” or “Non-Brand” campaign. Set this campaign to medium priority level in the campaign settings.
2. Set Priorities For Non-Brand
Now that you have your brand campaign set up, head over to the settings for your non-brand campaign. Make sure to set the priority level to high. Well, really the important thing is that non-brand has a higher priority than brand. So another perfectly viable option is to have brand set to low priority, and non-brand set to medium or high priority. This might all sound a bit confusing so far, why would non-brand be higher priority than Brand? Well, wait for it, this next step explains all…
3. Add Brand Keywords As Negative Match to Non-Brand
Go into your non-brand campaign and add in your brand terms and all associated keywords (plurals, misspellings, etc) into the negative keyword list for the whole campaign.
And there you have it! You now have a fully functioning brand / non-brand split in your Shopping campaigns. When someone searches for a brand term that triggers Shopping ads, they will not see the non-brand ads because those terms have been negative matched from your non-brand campaign. When someone searches for a non-brand term that triggers Shopping ads, the higher priority on the non-brand campaign will kick in and ensure they see the non-brand ad not the brand one.
Now you’ll be able to set specific bids for brand Shopping and split it out accurately on your reporting, just like in standard Search campaigns.
Mistake 3: Insufficient/Infrequent Bid Optimisation
How often are you currently updating the CPC bids on your Shopping ads? If you are like most advertisers whose accounts we review, the answer is, “Oooh maybe once a month, at most, but often a lot less than that.” This attitude towards bid optimisation, and a poor strategy behind bid optimisation, is probably the main reason why we see client Shopping campaigns underperforming versus their targets.
The Search advertising marketplace for your products is shifting, dynamically, ALL THE TIME. Competitor’s are adjusting their prices and releasing new products or updates. Search volumes are going up and down. Conversion rates are twitching according to time of day, day or week, and season.
Uncountable other factors are influencing consumers to buy more of or less of your products. This is happening in real-time, potentially thousands of times per day. With this in mind, you can’t really expect to have ideal CPC bids in Shopping if you only update them once per month, right?
There’s a challenge here, of course. We’d all like to have CPC bids updated in real-time to maximise our profit or ROI goals. But, what if you have thousands of products listed? How are you going to optimise so many bids with the frequency of updates required for optimal profitability?
The answer is, trust in the machine. Automated bid management by Google and by third-parties has gotten really good over the past few years. We are at a point now where, assuming you have enough conversions (about 30-50 conversions per performance group or portfolio will normally do it), an automated bid management system can do a very good job of dynamically updating bids and hitting your performance targets. The only remaining challenge now is setting the right Adwords goals (ROAS, CPA, Clicks, etc) to match your overall business goals (profit, revenue, growth, etc).
At Big Flare we’re a big fan of Adspert’s bid management tool. Despite not having access to as much data as Google’s free built-in bid optimisation tool, Adspert has often outperformed Google in many of our tests. However, as with any third-party tool worth using, Adspert is not free. Google also has it’s built in bid management algorithms that you can use for free and that do a pretty good job as well.
Whatever the tool or algorithm you use, we recommend you get on one right away. Shift your bid management system from checking once in awhile and updating manually, to setting up automated bid management and then checking on the results once per week to make sure your Adwords bidding goals and results are delivering the business goals you are trying to achieve.
To set up an automated bid strategy with Google, simply head over to: Shared library > Bid Strategies and create a new strategy.
For Shopping campaigns, and for eCommerce PPC in general, Return On Ad Spend (ROAS) based bid strategies are recommended. Out of the bid strategies available on Adwords, only ROAS can bid based on the actual return you are receiving, rather than just on the number of conversions or clicks.
Name your bid strategy something meaningful. If this is going to be a Shopping only bid strategy, then name it with “Shopping” at the start then add your performance goal too. Select which campaigns you want to include. This can also be done via campaign settings. Set your performance target (in the example above, a ROAS target of 300% was set) and then click save.
There is the advanced option available to set a minimum or maximum bid, but for most advertisers this is not recommended. Simply let Google bid whatever it needs in order to achieve the target you set. As long as your performance target is being met, who cares what maximum or minimum CPC bid is being used? Click save and you are good to go.
A Note on Not Panicking
AdWords bid strategies take time to learn and adapt. Do not just switch it off in a panic after two days if it hasn’t met your goals yet! Depending on how many conversions you have, the bid strategy could take anywhere from 1-4 weeks to reach optimal performance. When you start out, be patient and wait for it to dial in your bids. If you drill down into the bid strategy, you’ll be able to check the progress of its learning curve.
And there you have it. The three most common AdWords Shopping campaign problems we see in client accounts, and the exact solutions for them. There are, of course, plenty of other common problems that we see coming up in more than a few Shopping campaigns, so this is not the be all and end all of Shopping campaign improvement. But, if you are currently running Shopping ads in-house, it’s likely that the three above items are three things you could be doing better with only a minimal investment of your time and effort.
Are you making any of these mistakes and did this article help? Are there other challenges you are having with Shopping? Feel free to share in the comments below!
In some cases, PPC doesn’t pay off from the very beginning! Don’t just give up without giving it another try. Follow these simple yet powerful tips to improve your AdWords performance.
1. Use Conditional Keywords
Are you using carefully selected keywords but they are not performing well? You need to switch your keywords strategy right away. Sometimes even targeted keywords don’t convert (for multiple reasons). For example, information seekers search the same keywords and buyers, but are not ready to buy.
Let’s suppose you are a house removal company in New York and targeting the following keywords…
- house removals
- house movers
- moving companies
- removals company
They are your main targeted keywords and are eating your most of the budget. Pause them! Instead use conditional keywords where the user is looking exactly for the service you are offering but with a solid and well defined condition. For example:
- international removals
- interstate moving companies
- furniture removalists
- nationwide moving companies
- apartment movers
- office relocation
- local moving services
In my experience, these types of users are most likely to convert at higher conversion rates. It will lift up your conversions and decrease your costs.
2. Use Long Tail Keywords
Long tail keywords are another great way to optimize your AdWords account performance. Don’t confuse them with ‘conditional keywords’.
As I said earlier, conditional keywords have clear and solid condition attached to the search query, whereas long tail keywords don’t necessarily need to have one.
For example, for removal companies, long tail keywords might be:
- best house removal companies
- cheap house movers in new york
- where can I find moving companies
To get even better results, use long tail conditional keywords which in this case could be:
- long distance interstate moving companies
- discount interstate moving companies
- best interstate moving companies nyc
- moving companies for interstate moves
- full service interstate moving companies
- interstate moving companies near me
One conditional keyword will have several long tail keywords so there is a solid chance that you get good conversions from such keywords.
3. Create “Near Me” Campaign
This is, perhaps, the most neglected technique for getting more conversions and improving account performance. People are getting more localized day by day in this mobile world.
More than 50% of traffic is already on mobile and they are well aware of the fact that Google knows their location already. I know if I type “near me” next to my search query, this will bring up much more localized service providers.
But there is a catch. In this scenario, you are targeting entire New York and in such big cities, not everyone is close enough to visit your business. Even if they see your ad, they are least likely to convert. You need to target those who are physically close to your business location when make a search online.
Here is how you do it—create a separate campaign for “Near Me” searchers. Let’s suppose you are based in Long Island. Go to you campaign’s setting tab and then location. Click on “Add Location” and then select “Near Me” campaign. Search for “Long Island” and add it.
Now you are targeting people physically located in Long Island (add as many nearby areas as you want). Add all relevant keywords. Few examples are below…
- house movers near me
- house removals nearby
- closest house movers
- local house removal companies
If, in some cases, you don’t find your area in the AdWords location targeting interface, try your zip code instead, and if doesn’t work either, target by “Radius”.
You can even target generic and broad keywords in this campaign with even less fear of losing money. Here are a few more tips to consider:
- Add localized messaging in your ad copy
- Use the location extension
- Enable the phone calls extension
There is another way to target visitors near your location. You must have your business registered with Google My Business. The next step is to link your “My Business” location with AdWords account. Read this Google guide for further instructions on how to link it.
After it’s done, heads toward the advanced location targeting window and click on “location groups” tab and select My Locations from the dropdown as shown in the image below. Select the radius you want to target, let’s say 2 miles and click on Add.
This is a convenient way to target users near your business location especially when you have multiple business locations. This targeting technique will apply to all business locations you have in your Google My Business account.
That’s how your ad will look like on mobile.
(Image courtesy of Google)
At this point you may want to add multiple radius targets to single campaign to find the best radius that works for you. Let’s suppose you targeted 2 miles initially. Now repeat the whole process again and target 4 miles, then 7 and 10 (or what makes sense to you). In this way, you can learn about the impact of distance from your business location and later set your bids accordingly.
You should also check the ‘distance report’ more frequently to see how far the customer was when she saw your ad. This can be found under the “Dimension” tab. Choose ‘Distance’ from the ‘View’ drop down.
(Image courtesy of Google)
If a user happens to be in the area close to your multiple business locations, he will see all of them in the ad and can choose one which suits him best.
(Image courtesy of Google)
Tip: Schedule your “Near Me” campaign to run during your operating hours.
Note: This technique is applicable to all other campaigns in your account as well and works perfectly fine for the visitors physically close to your location.
4. Enable Mobile Click-to-Call
This is most important and least used feature by AdWords advertisers in this mobile world. Please note that I am not talking about the “phone calls extension,” I am not even talking about the “calls from website.” I am specifically talking about “phone number click” on your mobile site.
There are 3 types of call conversions you can track from within AdWords interface. If you have a phone number displayed on your website and are not tracking all three of them, your stats are not accurate (especially when you have to report to client about AdWords performance).
i) Call Extension
Majority of the advertisers are well aware of this feature known as “Call Extensions”. Visitors see “call” button next to your ads on mobiles or your “phone number” on desktop devices.
(Image courtesy of Google)
ii) Calls from website
In this type of conversion, Google displays a dynamically-generated forwarding number when someone visits your website after clicking your ad. Your own phone number will be replaced by this forwarding number, so you can measure how many people called you after landing on your website either on mobile or desktop.
(Image courtesy of Google)
Fewer AdWords advertisers are taking advantage of this feature as compare to call extensions.
iii) Phone click on mobile site
This is perhaps the most important call tracking source. It will turn your dead phone number on your mobile site into live clickable phone number. Here is how Google describes it…
“When someone visits your mobile website after clicking one of your AdWords ads, conversion tracking can help you identify clicks on your phone number. Unlike website call conversions, this feature only tracks clicks on your phone number, not actual phone calls. You’ll see click data instead of call data (such as call length) in your conversion reports.You can measure clicks on a text link, image, or button.”
(Image courtesy of Google)
When visitors land on your mobile website after clicking your ad and decide to call instead of signup, they have to manually dial your number. Giving them the option to click to call will definitely increase your conversion rate.
On a side note, I checked ten different websites currently running paid campaigns in New York for “house removals” related keywords and found the following worrisome facts…
- 3 out of 10 companies don’t even have a phone number displayed on their websites
- 4 out of 10 websites are not even responsive (mobile-friendly)
- 10 out of 10 are not tracking phone number clicks on mobile devices
- 9 out of 10 websites aren’t tracking calls from website
- 8 out of 10 haven’t enabled click to call function for organic users on mobiles
- 2 out of 10 don’t have Google Analytics installed on the websites
- 9 out of 10 don’t have AdWords conversion tracking installed on their websites
Let me remind you once again, all these 10 websites are currently running paid campaigns on AdWords.
5. Find & Use a Winning Landing Page
Easier said than done, right? Give me few moments and I’ll show you how easy it is to implement a winning landing page on your website especially for paid campaigns right from the beginning.
The best practice to be always testing your landing pages and ultimately find best performer. But it’s time consuming and can cost you good amount of money if you are running paid campaigns so where should you start?
Start from spying on your competitors, we need to take help from a PPC spying tool here. Let me give you an example of SpyFu which is very useful to look at the ad history of our competitors and for how long they are using the current landing page.
Enter a domain of your well established competitor and look back into history for how long they are using the current landing page, also check their ad copy, call to action, display URL, landing page content and If the competitor is using the same landing page for several weeks or months then you find yourself a winner.
6. Bid on Competitors
This might be the most effective and cheapest way to get more conversions. Most of the advertisers are aware of this technique and they are bidding on the brand names of their competitors but you know what?
They are bidding on their online competitors, what they are not doing is considering the offline competitors into online landscape.
Every business has offline competitors spending heavily on TV, print, or radio advertisement and you know what the interesting part is? They don’t even have a websites!
This is a huge opportunity for you to bring their customers to your own website.
Let’s suppose you just find out that you need a plumber and at that exact moment, you watch a compelling TV ad about “Acme Plumbing Company,” after the ad is gone, you don’t remember the phone number they shown on TV or the address, what you remember is—the company name.
So what the next thing will you do? You’ll search for them online.
I am getting 400 leads a month for one of my client by using this technique and these 400 leads are coming from a single competitor who is advertising heavily on TV and print but doesn’t have a website. I am just bidding on their brand name and you can imagine the cost/conversions, right? It’s less than 20% of avg. cost/conversion of that account.
Now it’s up to you!
Let me know in comments which technique you are NOT using yet?
Last week, Google announced a new change that will affect exact match keyword targeting. In the past, exact match meant exact match. Then, in 2012, Google announced a fuzzy match algorithm called “close variants,” which was supposed to capture plurals, misspellings, typos, and other versions of exact and phrase match keywords. With their latest update, the exact match targeting is getting bundled with the close variant targeting algorithm, which now ignores word order and function order. In this post, we’ll cover exactly what’s changing with the switch from exact match targeting, what you can do to mitigate any surprises on your end, and what experts are saying this means for the PPC world.
Rewording and reordering from close variant targeting will now include exact match targeting in AdWords. Close variant targeting not only ignores plurals, typos, abbreviations, and adverbs, but will also be broadened to ignore word order and function words. This means ads may be delivered when queries use a different word order or function words. Sadly, this limits the amount of control advertisers have over ad delivery, and further dilutes exact match targeting.
How this Update Affects Advertisers
The Google philosophy is as follows: capture as much traffic as possible with a wider net, then filter out what you don’t want, rather than building a smaller net that might not be big enough to catch everything you want. They’d prefer you spend money on some bad keywords than risk missing out on some potentially good ones.
Because of this, Google removed the ability for advertisers to opt out out of close variants in 2014. This means advertisers are being forced to place more trust on Google’s machine learning algorithms. This is obviously a smart way for Google to make more ad revenue, as they’ve claimed that early tests show advertisers could see an average of 3 percent more exact match clicks while maintaining similar click-through and conversion rates.
Below we’ll walk you through how the close variants are changing:
Function Words Could Be Ignored, Changed, or Added
Function words are essentially words that don’t have meaning on their own within a search query. Google defines function words as prepositions (in, to), conjunctions (for, but), articles (a, the) and other words that usually do not change the intent of a query. With this update, exact match could ignore, add, or change these function words to match with similar queries.
However, Google specifically states that the function words will only be ignored from the query when it does not change the meaning of the keyword. For example, “hotels in new york” can safely ignore the function word “in” because it doesn’t change the meaning. However, in the keyword “flights to new york” the function word “to” would not be ignored, because a “flight from new york” is not the same as a “flight to new york.”
Below are more examples from Google:
In some cases, two keywords can share the same meaning, and when they do, Google may reorder the keyword in order to deliver your ad. It’s important to note that word reordering will never add words to your keywords or the search query. For example, “buy new cars” and “new cars buy” likely mean the same thing from an intent standpoint. Exact match will use that same logic to match ads with reordered variations of your keyword.
However, similar to the function words exception, Google claims your keywords will not be reordered to match with a query if it changes the original meaning of those keywords. For example, the keyword [SFO to JFK] will not match to the query “JFK to SFO” because the destination and search intent is different. Below are a few more examples of how words can be reordered when using exact match with close variants.
How to Reduce the Impact From These Updates
Let’s face it. There’s no way around this global change and you’re not going to stop advertising on Google because of it. Because this update puts the onus on advertisers to explicitly state what queries they don’t want their ads to show up on rather than the ones they do want, it is important that advertisers be more diligent than ever about digging through search query reports and thinking ahead to prevent unintended consequences when word order matters.
As the exact match targeting update is rolling out over the coming months, here are a few things you can do in that timeframe to reduce the impact these changes have on your campaigns:
- Update your scripts. If you are using a script like the one from BrainLabs to make exact match exact, it will need to be updated as so.
- Review all your existing exact match queries and determine if the loss of function words or a reordering of the words changes the meaning. If so, you’ll want to add those variations as negative keywords in your campaigns.
- Review close variants in your Search Query Reports to see if other variations are currently being triggered that could be affected with the update, and add those as negative keywords.
- Going forward, schedule more time in your day to mine through your Search Query Reports, especially for close variants (screenshot below)
The Expert Take
It’s been a week now since the announcement and the PPC community has been buzzing with feedback about the exact match targeting update. We asked some of the leaders in the search world for their thoughts:
“As control freaks, this update spells doom and gloom to our tidy and structured accounts. But as marketers, this could mean additional query growth and less “Low Search Volume” keyword status.
We all survived the mandatory opt-in of close variants back in 2014 and this update only strengthens the point that no keyword structure can lead to success without active management. In the meanwhile I plan to schedule out some more SQRs.”
“This change seems to be less about “increasing reach” and more about herding advertisers down the path toward keywords as a lower level targeting layer. You can’t accidentally add exact match terms to your account, you have to do it deliberately. And, what the change actually does is already covered in AdWords by using the broad match modified match type for target keywords. It’s clear that Google is viewing related queries (or “close variants”) as fully interchangeable, so if that does not work in your particular situation, it is going to require more vigilance and work to try to keep the majority of queries matching for your preferred phrasing.”
“Most accounts: Accounts that use combinations of exact and modified broad won’t see any more impressions. Their biggest issue will be year over year bidding or traffic moving from one ad group to another one in trending data.
The rare word order account: When word order or word variations matter, either in meaning or click values, then you need to watch the changes closely. If Google gets it right, then it might not be worth the effort to manage this closely. If Google gets the meaning wrong, then you will need to do a lot of work in order to properly manage which ad groups are receiving the clicks and their associated bids.
Small accounts using mostly exact match: These accounts are going to see the largest impact. If you are mostly relying on exact match with very little modified or broad; you will see an increase in impressions. Make sure these additional impressions are converting at the same rate as your previous ones.”
“At first I’m thinking this gives us less control. Google simply wants to push out more ads to make more money. When you dig into their statement around machine learning and think about all the data they have on how we search. I start to think if anyone can pull this off, it’s Google. The big “what about when” around this is how does this machine take into account sentiment and the true context of what someone searches. 20% of searches every day have never been done before and if that stat is still true…. what would the machine do in those cases? It has no reference for something that hasn’t happened yet.”
“Inexperienced businesses will be hurt the most because they don’t know all the intricacies of AdWords. But if you have an established account, with a long history of appropriate negatives, I believe you won’t see a huge impact unless word order matters. If I’m a Wisconsin company selling a cheese head, I don’t want to show up when someone is searching for head cheese. Look it up. Two WAY different products.”
“I’m not too concerned about the AdWords exact match update. For years, Google has been allowing close variants to show. As long as PPC Specialists have tightly themed ad groups and constantly review search query reports, these close variants can help efficiency. I don’t see the mentality changing with how exact match will now be triggered. As long as advertisers remain steadfast reviewing their queries, there shouldn’t be major issues.”
“With changes like this sometimes there’s a “the sky is falling” response and in the end it doesn’t turn out as bad as everyone thought. It’s still unclear how exactly it will affect performance, but it’s making us rethink some of our fundamental strategies around keyword, adgroup and campaign structure. I’m less confident in exact match campaigns and adgroups and more worried about what I’ll find in the search query report for exact match terms. The “golden years” of true exact match ended with close variants a couple years ago, and while this continues the trend of less control, there are still many many levers we can pull to optimize and improve our digital campaigns. I’m still optimistic about the PPC future.”
“I understand the reasoning for Google’s changes in AdWords, and think this is simply a logical progression from their previous close variants change in the Exact Match Type. Google is trying to make it easier for higher-intent terms to be bid on in advertisers’ accounts who haven’t taken the time to find and add every possible query into their account. Makes sense, right?
On the other hand, frankly, I and many others were already ensuring this was happening by including tightly controlled Broad Match Modified and/or Phrase keywords in our accounts to pick up these “exact close variants”. Thus for us, the change is redundant and unnecessary in our accounts (though admittedly it can now force these exact close variant matches out of lower bid BMM keywords and potentially provide better immediate bidding accuracy on those semantically matched terms in the future – as I have written here: Match Type Segmentation Wins with Google’s Exact March Matchness Update).
My thoughts are that this change will not kill the majority of accounts (I believe some high CPC, probably B2B, terms that have multiple meanings will be hit hard with this), I just think it’s one of those unnecessary changes that further dilutes the original purpose of the Exact Match Type. Exact means, well, “exactly the same” and this new change strips away that meaning even further.”
Have additional thoughts to share on the Exact Match Targeting update? Leave a comment below!
When you can promise new clients a 250% increase in PPC campaign growth in a year, you must have a pretty good idea of what you’re doing. Bryan Gaynor, a Digital Marketing Account Manager at Hanapin Marketing, does. He knows a lot about PPC and digital marketing in general, but we invited him onto The PPC Show to talk about one of the newest forms of advertising – programmatic. You can listen to the whole episode for more details on the following topics:
1. “Reach the right person at the right time with the right message.” There are many definitions of programmatic advertising, but Bryan’s simple explanation mentions piecing all available data together to create messaging that’s personal to the viewer. Other definitions include the automated aspect of programmatic, where algorithms are used to purchase ad space.
2. “There are 80-90 other sources of inventory outside Google Display Network.” Though many marketers rely solely on GDN, programmatic offers access to up to 90 other sources of inventory. Taking alternative routes not only helps diversify your strategy, it could also lower your bids and increase conversion rates.
3. “If you want to sell anything, you have to specialize in it.” While Bryan doesn’t see a need to form a new team around programmatic, he does encourage managers to start testing now to gain an understanding that’s well-developed enough to properly recommend programmatic to clients. The principles are similar to how you’d set up a search or social campaign, so getting started is self-serve, but there are nuances managers will need to become familiar with. DoubleClick, The Trade Desk, and Centro are great starter platforms.
4. “You can get results with just $500 per month.” Bryan warns smaller budgets will yield fewer data points, which makes interpreting results a bit more challenging, but the minimal price to experiment with programmatic is very reasonable.
5. “Testing programmatic starts with what you’ve already tried.” You don’t have to start at the beginning to test out programmatic. Start with where you currently are, and use more in-depth reporting to further dissect placement and audience, so you can fine-tune messaging.
6. “The level of reporting with programmatic gets very deep.” Programmatic’s deep and detailed reporting helps surface information that might not show up in the reports you’re pulling now. Bryan’s worked with clients that were able to drill all the way down to adjust creative for viewers seeing ads on older phones.
7. “If it’s too good to be true, it probably is.” Though much of programmatic is automatic, it’s still important to keep a close eye on reporting, especially exchange or network reports. Taking a look at data related to impressions, traffic, and conversions will help flag and filter instances where performance might be incorrectly affected by non-human interactions (bots). Reviewing reports on a weekly basis will ensure you’re not spending money on wasted inventory.
Want all of Bryan’s tips? Listen to the whole episode below:
Bryan will be speaking at Hero Conf April 18th to 20th in Los Angeles (with AdStage CEO, Sahil Jain), where he’ll hold a session on programmatic for PPC managers. You can follow him on Twitter to stay up to date. For even more PPC wisdom, check out Hanapin’s library of resources.