Retour haut de page
Who wouldn’t want Jarvis - the complete AI-based assistant Tony Stark (Ironman) uses in his home, lab and metal suit - to help them as they go about their business? Well, more and more of us are stepping into this brave new world with baby steps, beginning with assistants like Apple’s Siri and Microsoft’s Cortana.

Voice Search Technology

The popularity of these voice search-driven technologies has exploded in the past few years. Voice search is still a relatively new development -- although the technology has existed for a while, voice-first devices were still being perfected and introduced in 2014. By the end of the year, just 3 years later, there are set to be over 30 million devices in circulation. This exponential rise in popularity illustrates the impact of voice search and voice-first devices, as well as users’ fascination with them. Beyond this, voice search is integrated into computer operating systems - the latest versions from Microsoft and Apple have their respective voice guided assistants embedded in their UI. 72% of people, according to this survey by Fivesight Research, say they use Voice Assistants to supplement standard search engines. Particularly now, with physical devices designed principally to be driven by hands-free assistants like the Amazon Echo and Google Home, voice search and voice commands are being used much more often than in the past. Rather than only being relied on to make calls or search the web, voice search is now being used to help with simple, day-to-day tasks. You can control your home’s heat, lighting, music and appliances with voice. They’re on board TV streaming boxes, gaming machines like XBox and the range is growing all the time. Because of this, voice devices are becoming a big part of many individuals’ everyday lives.

Voice Search and Advertising

Because these tools are still in their infancy, there are no advertising options yet. This is complicated by the fact that rather than directing users to web pages, voice-first devices can, in many cases, provide users with an answer directly. It may be worth noting, by the way, that only Google’s voice search/assistant has its web search features powered by Google. The others - Amazon (Alexa), Siri and Cortana - are powered by Bing search. Google’s CEO, Sundar Pichai, recently wrote on their blog that the Google Assistant and subsequent Google Home were the result of advances in artificial intelligence, specifically in areas like natural language processing and voice recognition/translation. These developments allow users to interact with voice search devices more naturally, because “the assistant is conversational—an ongoing two-way dialogue between you and Google that understands your world and helps you get things done.” Rather than focusing solely on providing information, the Google assistant and similar devices operate with the goal of helping users with everyday tasks like playing music, making reservations, checking movie times or sports scores, getting directions or making calls. The unique way that users interact with this new technology is what makes voice search a new frontier, particularly in the realms of advertising and paid search.

The Future of Voice Search

None of the major players have commented on if and when they will present options to market to users. Amazon has a model for now of using Alexa to make ordering from Amazon easier (and more likely), but for the others, beyond simply selling dedicated hardware and adding value to existing hardware and software, there are no obvious ongoing returns they can utilise if they don’t eventually embrace advertising in some form. Currently, there are potentially “second hand” benefits of the likes of Google’s Assistant in that they can plug into apps businesses provide to book or order from them. But often, this is not an open plain for any business who wants to do it - although Alexa is leaning more open source than the rest. Right now, such plug-ins are anything but easy and intuitive and a better way needs to be worked out so that the boundaries of these voice assistants are not so apparent as they are now. Because these devices are still so new, their rapid rise in prevalence raises many questions: Who is using voice search, and why? What opportunities will there be for advertisers to respond to voice search? How may it change the search industry? What could be the future developments? When will I be able to order a flying metal exo-skin armed to the teeth and giving me superhuman strength? While it is difficult to predict exactly how voice search will continue to grow, it is clear that it will eventually have a huge impact on the search field in the next couple of years. With richer, more customised results, voice search and voice assistants have the potential to revolutionise both how people use search technology and how advertisers and businesses can link up with them to serve their needs. There may be a longer wait for the metal suit, though. For updates on how advertisers are responding to voice search, tweet at us @ESV_Digital_UK or follow us on LinkedIn.

 

After drugs, fraud in digital advertising is one of the most lucrative illegal activities in the world.

When talking about digital fraud, the reaction is often to think about hackers and computer security issues. However, there is a lesser known digital fraud, but its financial impact is far greater: fraud in the field of digital advertising. The World Federation of Advertisers estimates that the impact could be up to 30 to 40 percent of global digital media investment in 2025, an amount that could amount to $ 150 billion.

Fraud in digital advertising is divided into five main themes:

  • The false traffic. Robots accounted for 6% to 9% of impressions recorded in 2015. (2)
  • The biased returns. Some publishers manipulate the format of ads to increase their returns, for example by forcing users to click on an advertisement in exchange for participating in a lottery or other material benefit.
  • The visibility. More than 40% of advertisements invoiced and broadcast on the Internet are not actually seen by visitors, in particular because they are placed below the page break or because they are loaded after the user has already left the page.
  • Poor targeting. Certain target settings, based on socio-demographic criteria or personal interests (age, geography, sex, occupation, hobbies ...), do not correspond to the reality of the Internet users actually exposed to the advertising campaign.
  • The safety of brands. More than 10% of ads are placed on sites with violent or pornographic content which any mainstream advertiser would want to avoid.

In order to assess the extent of the fraud of which the advertiser is a potential victim, the latter has two coupled approaches to evaluate the quality of its digital investments:

A purely quantitative approach to analysing and cross-analysing data available in analytics and allocation tools, ad servers, advertising auction platforms, and programmatic purchasing management tools (DSP) to detect metrics that do not match the expected values.

A more qualitative approach to analysing the implementation process of campaigns by the advertiser and its agencies (briefing, reporting, choice of KPIs, objectives, targets, etc.) in order to contextualise and understand the analysed data.

After the analysis phase, it’s time for action recommendations. As an example, here are some typical actions:

  • Access and store data for ongoing storage and analysis.
  • Integrate technological tools to analyse the media mix and traffic, while being vigilant on the relationship between performance and cost.
  • Define processes within the advertiser (management team, marketing, digital) and the agency (brief, debrief, reporting).

For each proposed recommendation, it is important to evaluate the resources (internal or external, tools, processes, etc.) to be mobilised by analysing the relationship between full costs and impact in order to prioritise the actions to be carried out. It is important to emphasise that these actions will not eliminate fraud, which is an unattainable objective, but rather to limit fraud and limit its volume.

A biased promise

It is paradoxical to think that the share of fraud is larger in digital advertising than in traditional advertising, even though digital advertising promises total transparency of media investments: on the advertising used, on the site where it is published, on the target user (via cookies), whether he has been exposed to the advertisement in question (and in theory, on the price paid under the Sapin law), and on the conversion measure for each purchased advertisement. But in reality, it is the infinite granularity and the related complexity of the purchase of space on the Internet that opens possibilities to those who would like to defraud advertising on the Net. This granularity is the primary strength of advertising on the internet, which makes it difficult to control all the advertisements purchased and their compliance with the orders of the advertiser.

Control and measurement tools to limit fraud in digital advertising exist, but they alone will not suffice. Success depends on the cooperation of all players in the industry: advertisers, agencies, space vendors and technology solutions providers. It is important for the future of the Internet ecosystem and to preserve the confidence of the actors who trust it to take the subject of fraud in hand.

To learn more about how fraud in digital advertising might have an effect on your company, tweet at us @ESV_Digital_UK or follow us on LinkedIn.

March Madness is here, but that doesn’t mean just basketball. This March has seen all kinds of controversy swirling regarding display advertising and how display ads were being shown on either videos (such as YouTube) or sites with unsavoury content. The timeline on this has been short and of surprise to no one heavily involved in Digital Marketing.

The Tip-Off

The whole whirlwind appears to have been kicked off by an article in a British newspaper, where reporters had found ads of various advertisers showing on terror-related pages and videos. Initially, the finger was being pointed - at least in part - at the advertisers and forced many to respond publicly at charges that they were (even inadvertently) funding criminals and those who support mass murder. However, this soon morphed into the blame being directed more towards the advertising networks - Google being easiest to name for journalists - and multiple commercial and governmental organisations announcing advertising boycotts.

Non-digital experts may feel like this is all very sensible and that there’s been some bad behaviour somewhere in the complicated world of online advertising but the reality is that this is “madness”.

Timeout

Display advertising, in fact, is nothing like the buying of ad space in a print publication, where you can negotiate sizing and positioning in fine detail, and where you can guarantee that anyone reading the page on which you have the ad will see that ad. It is in no way as direct a process as throwing an orange ball into the netted hoop of your choice!

Display ads load in real time on a page which could be totally dynamic (e.g. a forum), could also contain the content of multiple sites (embeds) and potentially every page load of this same page could show a different ad not only to every single visitor but even whenever the same visitor reloads the page. In addition, the uniquely digital ad format of remarketing cannot be discounted from reducing control for any given advertiser.

Remarketing is the delivery of ads targeting people not publications on the web based on their behaviour (on the site you are marketing, typically).

In addition, there is the concept of targeting users by the keywords they use on search engines and delivering display ads on whatever site they visit on the strength of those searches.

It is absolutely true that there is the ability for advertisers to exclude ad delivery from classes of sites, individual URLs, and even text content in pages but the latter options and the overwhelming approach on this is, inevitably, reactive and opt-out rather than opt-in. The point being that no advertiser is consciously “choosing” to advertise on objectionable page content.

Foul

So are advertisers right to blame the likes of Google instead and boycott them? Well, this does seem to be rather an over-reaction to ill-informed bad publicity in an effort to throw public opinion/outrage in a different direction, rather than truthfully reflect the complex nature of this challenge. However, Google can and should restrict more carefully ad delivery on their YouTube videos and even slight over-compensation is preferable to under-regulation.

No broad digital advertising network can possibly hope to guarantee 100% brand safety if they are to offer reach across the wider internet and not just highly established websites. But also, it’s about reaching people, not just websites, and it is impossible to control the sites people visit after being added to an automated remarketing list against which they will be shown ads.

Blocking shots

That said, if, as an advertiser, you are made gunshy by the unknown risk out there, there is a way to maximise your brand safety and that is to use the several specialist ad networks out there who only deliver to suites of established, commercial sites onto which objectionable content will not creep. The usual cost for this conservative approach is, however, a smaller reach than competitors and, often, higher CPMs/CPCs. This is because it is rare that those networks have exclusive access to those quality sites and so you are still competing against just about everyone else but reliant on that limited set of sites for all your traffic, unlike the competition.

Should a larger network, such as Google’s try to more pro-actively vet sites - and we’re not saying this shouldn’t happen, welcome improvements can always be made - this may serve to calm the situation down. They have announced plans and this is a good start. But businesses and other advertisers need to temper their expectations and not demand 100% perfect brand safe yet broad reaching environments in the realm of digital display advertising.

For more information on optimising your performance display advertising, Tweet us @ESV_Digital_UK or follow us on LinkedIn.

Today there are over 3.5 billion Google searches performed each day and around another 2 billion on Bing. Obviously, a lot of these searches are in no way going to provide useful traffic to your business and those who manage Search Engine Advertising try to use tools at their disposal to qualify clicks as far as possible (Ad messaging, the keywords that are bid on and the tightness of the match types used, amongst other things). But if you go too far in these efforts, it very possible to miss out on large amounts of searches and customers that are relevant to your business.

To ensure that you are reaching the largest amount of people that might be interested in what you have to offer, it’s important to consider the below growth options. They do often require more intensive management time to optimise but they can often be worth it.

For those who are not clear on Match types, this article will fill you in.
[full-image bg=”https://www.esvdigital.fr/wp-content/uploads/2017/02/google2.jpg”]

Broad Match

Broad Match runs your campaigns on searches that aren’t necessarily in your keywords list, but would still be broadly relevant to the product or service that you offer. This can be refined significantly by using the “+” modifier next to a term in your keyword – where this is used, the term tagged with this will have to be present in the search that is matched – “mens fluffy +slippers” will match to “cheap slippers” but not “mens fluffy sweaters”, for example.

So by judiciously using this match type you can hope to win some ad space against the 1 in 6 searches that have never been seen before. You can also use this traffic to fish for repeated searches you haven’t yet bid on to add to your keyword list.

Phrase Match

Phrase Match: the ad shows when the user searches your keyword phrase with extra terms before and after or a close variation of the phrase itself. This is tighter than Broad but way broader than Exact Match and, when managed right – excluding Exact terms or bidding less than Exact – you can still get some valuable new traffic.

In addition to ads that are based on the keywords, Dynamic Search Ads (DSAs) can create advertisements based on your website.

Dynamic Search Ads

Dynamic Search Ads use your website to tailor advertisements to potential customers who are searching relevant keywords. Headlines are written and landing pages are chosen from your site automatically in order to be relevant to the search terms made by a user in real time. DSAs do not involve bidding on keywords, instead you nominate a section or sections of your website.

To ensure you don’t take traffic you’re rather have come through your keywords and other campaigns, you should bid less on DSAs than the broad keywords. The great part of this is that you will get traffic you may not get elsewhere and is a great way to take advantage of peak traffic periods. Also, if you just don’t have the staff resources to look after a large product portfolio, DSAs can help get you some relevant ads without needing to construct and manage massive keyword collections.

To get the best out of DSAs, though, you do need to add lots of negative keywords and manage your bids carefully.

So these types of approach can get you some useful, cheap, and otherwise missed extra traffic and sales for you. The main thing to bear in mind, however, is that they all take a bit more manual nuancing to make them work and they won’t work perfectly (e.g. profitably) right out of the gate.

If you would like to hear more about how these advertisements can benefit your campaigns, tweet at us @ESV_Digital_UK or follow us on LinkedIn.

[box]

The success of metasearch engines such as Trivago, Kayak, Tripadvisor, hotelscombined, hipmunk and others has been well documented, with impressive YoY growth figures over the last 10yrs. The foundations of this success and popularity come from continued traveller vigour to search, compare and review as many hotels and holiday packages as possible to find the best real deal.

However, if you ask the average traveller if they have heard of Google Hotel Ads (previously Hotel Finder and Hotel Price Ads) then the overwhelming majority of travellers will provide you with a blank face. Yet, in fact since its launch in 2011, Google’s hotel specialised metasearch engine has slowly and steadily become one of the market leaders of the online travel sector.

[/box]

So what are Google Hotel Ads?

Hotel Ads are a series of room rates listed under the integrated google map block for both brand and non-brand hotel search queries, highlighted in yellow in Step 1.
step3
Once the traveller selects their desired hotel, Google provides another selection of hotel tariffs from both OLTAs and rival metasearch engines, see Step 2. These listings are then prioritised by Google via one of two methods: the lowest price or in the case of price parity the listing that has bid the highest cost per click (CPC).

Up until 2011 both Google maps and Universal Map Search blocks had experienced low paid advertising exposure, providing hoteliers with a cheap traffic alternative (organic search) to Adwords campaigns. Now, if hoteliers, OLTAs and rival Metasearch engines want to appear above the fold on a desktop or mobile queries, they either have the choice of creating a traditional Adwords campaign or connecting to Google’s Hotel Ad feed or both.

[box]

Can anyone join hotel price ads?

Yes and no, any hotelier or travel agent can connect, but you need to connect via Google’s partner list. The listed partners provide a go between for Google’s Hotel Ads API and a hotel’s daily rate data. In most case these partners https://www.google.com/intl/en/ads/hotels/find-a-partner/ are flexible in terms of tracking and are able to integrate all the major tracking solutions, including ESV’s own SEM and Multi-Channel platforms.

What will Google Metasearch next?

With the major Travel and Shopping sectors already integrated into the universal search interface, Google will be on the lookout for new paid revenue streams to finance Alphabet’s new ventures.

Possible targets could be financial services again after Google pulled the plug on Google Compare. One must remember Google shopping has had many stop-starts & evolutions since the days of Froogle. Other big sectors could be the automotive sector or property; in fact anything’s possible with Google even potentially personal services such as barbers, dry cleaning or key repair. One thing’s for sure, the commercialisation of the Google interface won’t stop until people start to stop using Google.

[/box]

[box]

If you’ve been having trouble hitting your PPC Key Performance Indicator (KPI) for the business, or if you are hitting it but you’re not sure why you’re not seeing growth, this could be because you’re working towards the wrong metric.

Getting the right measurement for your sales target together with the right top-level number can transform your performance.

[/box]

What efficiency KPI is best for you?

Here’s a basic outline of what options you have:

ROI: The amount of money you have made from your ads divided by the money you have spent on them. You come out with a ratio (e.g. $200,000 revenue ÷ $80,000 ad cost = ROI of 2.5)

Cost of Sale (CoS): Essentially the same equation as the above but reversed (spend divided by Revenue) and is a percentage number.

Gross Margin: The profit percentage made from selling a product or service after deducting expenses (e.g. manufacturing costs) often called “Cost of Goods Sold” or COGS before you deduct the cost of ads. This can be used in lieu of outright revenue in the above two metrics for a more restrictive (but technically more accurate from a bottom line perspective) efficiency figure.

Cost-Per-Action/Acquisition (CPA): The average amount of money spent on advertising needed to see a conversion (or specific action).

Cost-Per-Lead (CPL): The average amount of money spent on advertising in order to generate a new lead.

Lifetime Value (LTV): The specifics of this can be calculated in a number of ways but it essentially is the financial value of a user over the long-term, not just the first conversion. The customization of this number can be through if you want to factor in costs, the timeframe you choose to use and whether there is a fixed or ongoing period of relationship with the customer. This too can be used as a substitute for Revenue in the efficiency metrics.

[box]

Decision

For many businesses, it’s better not to just look at one number. If you do, you risk harming yourself in another metric. For instance, beating a CPA target is easy, you can do that in a heartbeat but doing it this way will mean you reduce your total volume. The hard part – the bit that draws on expertise – is to both reduce CPA and grow conversion volumes (or revenue).

The other part of choosing the efficiency metric, is hugely dependent on what your business is, it’s business model and the pattern of its profitability. In other words, ask yourself:

  • Do I sell a small or large range of products?
  • Are the price points across the products a wide or narrow spectrum?
  • Are the expenses I incur hugely variable from product to product?
  • How likely is repeat business?
  • How long is the buying decision process?
  • How well known is my brand?
  • How predictable is my profit margin?

The more static your margin numbers and product range, the more suitable the likes of CPA will be. The more variable the Average Order Value (AOV) or customer lifetime length could be, the more likely ROI/CoS will make sense. But even then, a single, top-level number might distort your account development and it may be better to stick to product category-level (or your business’s equivalent) efficiency targets instead.

If you are going to go for an ROI or CoS target, then ideally you will not have a marketing budget limit. Profitable is profitable and if you’re beating ROI targets, then by definition, you’re making money on whatever you have spent, be it $100 or $1m.

[/box][full-image bg=”https://www.esvdigital.fr/wp-content/uploads/2016/08/performance.jpg”]

Analysis

So you need to weigh up what level of efficiency metric you want to use as a target but which also does not need to compromise your growth potential. If your PPC account has been running for long enough, you might just have the data to identify this magic nexus.

Use as many data points as you can and set up a scatter chart of Revenue (or gross margin) by CPA/ROI/CoS/ROI/CPL and see what patterns you get. You will probably want to look at Spend vs efficiency in the same way for other reasons.

It’s possible you’ll see a place where revenue plateaus or peaks at a range of the efficiency metric before you see diminishing returns. But beyond this, you can use the data you have to see what the increment costs are of getting incremental revenue/sales by using your real world data and Google’s budget and bid estimation tools.

[box]

Careful

However, there is another approach to metrics that is too easily overlooked. The volume can grow or be fine but the efficiency is worse – but the account is not actually performing worse. We explain this little conundrum in this blog post.

Knowing this could also help you choose target numbers as well as the KPI you want to aim for.

Do you have more questions about how to choose the correct efficiency KPI for you? Tweet at us @ESV_Digital_UK or follow us on LinkedIn.

[/box]

A PPC account that is running pretty well can always do better. But why rock the boat? What else is there to do other than optimise what you’ve got?

I’m here to tell you that if you don’t have at least 10% of your PPC budget free to experiment at all times, you’re soon going to fall behind.

Here is how to experiment and move your account forward:

– Plan – not more than 2 quarters ahead because the landscape of features and ad formats changes about every 6 months.
– Try not to test multiple things at once otherwise it becomes impossible to unpick what test is behind any changes in performance.
– Do the math to determine how much data you need for statistically significant results.
– Google Campaign Experiments helps you try out new things in a controlled and scientific way.
– Be clear about what you’re hoping to see from a test – better CTR, lower CPC, better conversion rate, higher AOV etc.

You really do need to test, and the simple reason why is that your competitors certainly will be. So why let them have the edge? When it comes to digital marketing, you can’t be sure of anything until you have the data, and only testing will give you that data.

Good luck!

If you would like to hear more discussions on the importance of ad testing, tweet us @ESV_Digital_UK or follow us on LinkedIn.

The Google Display Network (GDN) is a continuously growing and changing beast, much like the Google Search Network, and it is very easy to be spending too much or suffering from too little efficiency. Here’s how to take control of how your GDN investment is spent with a bias towards making more efficient conversions.

Remarketing

There are still accounts out there that, believe it or not, do not utilize remarketing features. Remarketing is where every visitor is tagged with a persistent cookie and can be added to one or more remarketing audiences (set up in AdWords) so they can be targeted with dedicated bids and/or messaging.

For retailers, the likes of dynamic remarketing can be especially powerful.

If you’re not doing remarketing yet, you should start. Remarketing ads have significantly higher conversion rates (and better efficiency) than standard GDN ads.

Similar Audiences

Building upon the remarketing, you can broaden this out to Similar Audiences which give you a way to take all those millions of users and winkle out those who share a common profile with your website’s visitors.

[box]

Structure

The fact is that text ads and differently shaped banner ads can show on different (but overlapping) subsets of websites and, indeed, positions on those websites. This being the case, throwing all your ad formats into a single ad group is not only a recipe for lower-than-possible efficiency but also difficulty in managing and observing performance, period.

If at all possible, you should have one banner ad format per ad group – each ad group for each ad can start with the same settings, such as targeting, but evolve as the data dictates. This means you have much more granular bidding, targeting and reporting (just report at the ad group level) and much easier access to ad-level stats.

In addition, you should have a range of GDN campaigns that represent a spectrum of audience breadth. You want campaigns that cover the whole network (subject to the exclusions mentioned later in this post) and with targeted placements in other campaigns excluded plus any proven poor performing placements also excluded over time, then contextually targeted campaigns (keywords only), then placement targeted and so on all the way through to segmented remarketing and customer match (the ultimate remarketing).

Separate campaigns like this allow for you to assign budgets to control spend levels and get the bidding more in line with the relative performance of all these campaigns.

Exclusions

In the vast majority of contexts, it’s basically always worthwhile to exclude all the available exclusions from your GDN campaign. The method of setting these exclusions is described in the latter link.

In short, they stop your ad displaying on places that can be havens for robotic clicks and should save you some serious money. In addition, there is an experimental AdMob exclusion option. The other way to exclude app ad impressions – and this part of the GDN is often the weakest – you should use placement exclusions.

Placement targeting and exclusions

If you’re running placement-targeted campaigns in the account, you should take full advantage of this knowledge that those ads are serving on website X and try some tests.

  • Mention the website in the ad.
  • Mention a special deal for visitors to this site.
  • If the site is likely to appeal to a certain demographic, play directly to them in the creative.

Ensure you use placement exclusions for websites or placements that drain resources with no return. It’s the GDN version of a search query report.

[/box]

Device delineation

Device performance is going to very a lot, you’ll most likely find mobile devices to be the weakest so assess if it’s worth it to be there. But also, if you think it is, try to structure your account to cater to these devices with banner sizes and calls to action.

Qualify and stand out

Low conversion rates on the GDN may not have anything to do with the placement or other factors but actually be due to the ad creative not qualifying users properly. Make sure a user is as informed as possible about your offering (and limitations, such as pricing – if it’s expensive – or eligibility – “home owners only”) before they click and visit.

In addition, attracting the right clicks can be about standing out and getting people who are not actively hunting for your product or service in the moment to click on your ad. The most likely thing they’ll want to do is check you out as opposed to buy there and then so the call to action should be in that vein (“Learn more”, “Don’t miss out”) – they want to understand your brand and be warmed up to you first.

If you know that your clientele is generally men but the site you are delivering ads on is read by both genders, set up the ad creative to appeal more to men.

Conclusion

Not necessarily every tactic here will work for your specific account – that’s why you test – but hopefully it will give you some areas to look at and understand if there’s more you can be doing to drive GDN efficiencies.

If you would like to learn more about how you can take control of your GDN costs, tweet at us @ESV_Digital_UK or follow us on LinkedIn.

Your PPC account has been running for some time and you’re making a profit OK, but it’s always worth asking yourself if everything is really set up in such a way that there isn’t a significant efficiency leak.
We’ll lay out here how to check for some common, but easily missed causes for an account to secretly have much less efficiency that you have potential to actually achieve.

1 – Traffic control

If every time you push spend higher, the efficiency gets much worse (that would metrics like ROI and CPA), because sales don’t grow at anything like the pace of the spend, this is typically due to a problem with traffic control.

Traffic control is affected by 2 main elements:

  • Match type share of traffic
  • Intra-account cannibalization

To accurately get a feel for the true level of Exact, Phrase and Broad matching traffic, run a Search Query report and include the match type column. Count the share of traffic for each match type, if you have a minority of traffic coming through Exact match, you have a traffic control problem. You’ll need to grow exact match traffic by taking relevant and converting search queries and adding them to your account as Exact match keywords.

Cannibalization is simply where a search with which you want to trigger ad group A is inadvertently triggering ad group B. This calls for negative keywords and/or a clamp down on match types.

2 – Search Partners

What is the Search Partner Network? It’s all the non-Google sites that partner with Google to feature both on site search and web search. Often these are newspaper websites and other publisher but really it’s a range of different websites that is ever-changing.

Since ads are often displayed differently on these sites than on a Google property – different sets of ad extensions and differing numbers of ads – and because the context is not so clean as a Google SERP, the performance can and does differ on this delivery channel vs the Google Network.

Sometimes, the performance can be better but sometimes it can be substantially worse. And more than this, it can vary a surprising amount from campaign to campaign.

Whether you opt in or out of Google Search Partners is chosen at the campaign level and you should decide based on the long-term (3 months minimum) performance differential.

3 – Sitelinks

Sitelinks seem like an easy thing to get right and are often overlooked but if you do them wrong by taking traffic to bad pages that take users away from their buying route, you’re going to get some sort of hit on your efficiency.

Also, sitelinks are often not refreshed very much and you may find a dreadfully poor conversion rate on one that you should have replaced or paused a long time ago. Look at your sitelinks performance and see if there is a problem there.

4 – Falling Impression Share

It’s absolutely possible that you could see rising volumes yet falling Impression Share. Impression Share (IS) is your share of all the impressions your account was eligible for expressed as a percentage. So if you were eligible for 100,000 impressions but got 80,000, that’s an 80% IS.

On a top level, review the trend for IS and, in particular, Exact Match Impression Share. If it is falling, assess the campaign from top to bottom to understand why. Use the Auction Insights report to help too and use device segmenting to get a proper picture of the weaknesses.

Keywords that have always had rank of 4 and below will be easy to recover with higher bids, provided they have a decent Quality Score but you will need to keep a close eye them to see if they ultimately justify the higher bids.

As an aside, however, it’s always better to be losing IS due to Rank than IS due to budget. To fix the latter, either raise budgets or reduce bids.

5 – Device modifiers

You are probably aware that you can bid different amounts for Mobile Phones via upping or lowering your desktop/tablet bid by a custom percentage. However, you’ll soon be able to vary your bids by tablet, desktop and mobile.

This open an opportunity in particular to get a grip on Tablet performance, since it can often be worse than Desktop.

6 – Geography

In another example of where digging deeper can save you some unprofitable spend is running a geographical report. Analyse this data and you could find a part of the country that is draining money without returning profits. You can exclude such locations or you can simply vary your bids (by percentage) based on location so you can drive efficiency.

7 – Scheduling

Although we’re increasingly living in a 24/7 world, it’s still possible that your business doesn’t require 24/7 search ads. Ultimately, unless there’s an obvious reason that you shouldn’t run ads all the time, let the data tell you when to have live ads.

Look not only at days of the week – and, who knows, maybe there’s a totally dead day with no sign of life but, short of that, there could be clear days on which to bid less or more – but also hourly patterns within each day. This will guide to lower and raise bids and even stop ads for parts of the day that simply don’t work. It’s more likely, however, that you will nuance bids as opposed to pausing delivery.

Again, like all posts of this nature, this covers only some of the more common areas of budget or performance drain but it should help you checklist your account and help you build it up in a profitable way.

Need additional help identifying problems with your PPC? Tweet at us @ESV_Digital_UK or follow us on LinkedIn.