Category Archives for "Google Ads"
I was telling you last time about some of my early Google Ads clients, and the challenges I faced.
Let me tell you about one particular nightmare client…
Nightmare clients are rarely bad people, otherwise they would never become clients in the first place. But they come with little warning bells that jingle away, just audible if you know what you’re listening for. Those warning bells are desperately easy to ignore, especially when you need the money. Sometimes you only hear them in retrospect.
This particular client had sat on my email list for a while. Which is a good sign – not a warning bell! One day he contacted me, and asked how my services worked. We’ll call him ‘George’ (not his real name, obviously).
George’s company was using AdWords to generate leads for an offline sales process. The company was spending £25,000 a month on AdWords (about $30,000). Which for me was a large account.
George mentioned on the phone they had ‘been through’ quite a few AdWords agencies. But he’d like to ‘see what I could do’.
(Ding ding ding! Can you hear the warning bells jangling?)
On the eve of the project George sent me an email in badly-formatted English. (Clients who can’t spell is another warning bell.) ‘I’d like a daily report please Rob’, the email said. ‘The report should show yesterday’s spend by campaign, click through rate, conversions and cost per conversion.’
I never send daily reports to clients because daily fluctuations are misleading, and clients who obsess about them often fly too close to the sun financially. Weekly numbers are more meaningful. And monthly numbers hardly ever lie. Still, I ignored the screaming messages from my gut, and agreed.
I would always know if conversions were down for the day, because at 3PM George would email me. ‘URGENT Leads down Rob. Please action ASAP.’
I would rage internally at this. Please action what, exactly? I’m not a magician. Most of the time I would ignore it and rely on a natural upturn in leads the following day.
I’ve also since learnt that clients who routinely send emails with URGENT in the subject line, signed off with ‘ASAP’ are not great clients to work with. It’s a small clue that they see you as a vendor to be used, not an expert to be consulted.
Not long after the project had started I made a bunch of changes to the website, and George blew his lid. “We make changes on a month by month basis,” he informed me. “That way we compare apples to apples.”
Apples to apples – yeah right. You can’t have it all ways up. You can’t scream and shout about your AdWords results, but also refuse to make any changes to the website.
The project fizzled out after a few months. I was glad, too. The hassle wasn’t worth the management fee they paid me. Or so I thought.
One year later George got back in touch. His last business had tanked, and now he had a new business in another highly competitive market.
George asked if I would consider managing the account again, since I had ‘done such a good job last time’. I wavered for a moment. The warning jangles were all there. I knew George was a well-meaning but troublesome client. But I also needed the revenue.
So we went again. This time George wasn’t my main point of contact; he had a marketing manager, Jane*. Jane was nice enough, but clearly under huge pressure to perform. Once again I failed to set the boundaries on when I could and could not be contacted.
At 9.50 every morning my phone would go (Jane started work at 10). Despite my repeated warnings about daily statistics, she would want to discuss yesterday’s numbers and know ‘what I was doing for them today’.
“Nothing,” was the response I should have given, “because I’m not on payroll.”
I didn’t have the gall to say that at the time.
One week we had a spat about ad creation. I had created some new ads in the account, and some hadn’t performed as well as the control ads. “No more duds please Rob,” was the message I got.
No more duds. If you never write an ad that fails, you’re not really writing any ads. The longer I’ve done this, the more I’ve realised how hopeless it is to predict a winner.
A few weeks later Linzi and I went to Italy for a week. I tried to reassure Jane that her AdWords results were unlikely to tank for a few days without my daily hand-holding. I had no phone signal in Italy, and no computer.
Two days into our holiday I logged on to Wi-Fi in my hotel, and my phone buzzed to life. WhatsApp. It was Jane. “URGENT: NO conversions yesterday. George won’t stand for it. Please look at this ASAP.”
Really Jane? You’re going to harass me on holiday… by WhatsApp?
After three months of conflict our second project came to an end. George finally discovered that I had been ruining his monthly ‘apples to apples’ comparison by sending traffic to pages other than the homepage.
The big problem however wasn’t with George. The big problem was with me, because I hadn’t been clear enough about exactly who I was trying to work with. I hadn’t set the boundaries. I had heard the warning bells – twice – and I had ignored them.
There are numerous lessons here. George wasn’t getting the results he wanted because he was agency hopping and trying to dictate the solution. He was also deeply reliant on Google Search for leads. If leads dropped off one day – for whatever reason – everyone got it in the neck.
I truly believe that Google Ads is an unbelievable system, but there has to be some room to fail. Fail small and fail often, but improve overall.
*Also not Jane’s real name, for obvious reasons.
Do you remember the first Google Ads campaign you ever created?
I created my first campaign in 2007. I was working as a marketing intern at a software company near London. The company already had decent organic rankings for terms like ‘CRM software’ and ‘CRM solution’.
Those rankings did them little good, because the website itself was terrible. But they did manage to rank organically for some impressive keywords.
So when I approached the MD with a proposal for running paid ads, he was understandably sceptical…
Why would we pay for traffic, he reasoned, when we already had free traffic coming in? But the reality was we needed more leads. So eventually I solicited a £1000 per month budget to experiment with.
At that point I had known about Google AdWords for two years. I started following AdWords expert Perry Marshall in 2005, through one of Ken McCarthy’s System Seminar recordings. “Jump on in,” was Perry’s message, “cuz these cheap clicks are gonna dry up!”
The two main PPC book at the time were Perry’s Ultimate Guide to Google AdWords, and Howie Jacobson’s AdWords for Dummies. Later on I discovered Advanced Google AdWords, by Brad Geddes. I read all of these, and was well versed in the principles.
Of course, the clicks didn’t feel cheap at the time. When Google first introduced AdWords in 2002, you could buy clicks on any keywords for pennies. 10 cent clicks were commonplace. By 2007, we were paying between £2 and £3 for CRM-related terms. Today, those same clicks would cost at least £15.
For two years I had studied the principles of Google AdWords, without really doing anything. I’m more of an accidental entrepreneur than a serial one. I didn’t have a business at 21. I was a studious and slightly apprehensive marketing intern, desperate to learn as much as possible.
Then I created an AdWords account for the CRM company, and realised I had no real idea what I was doing…
I knew about keyword match types. I knew to avoid broad match keywords. But how many keywords were you supposed to put in an ad group? Should you put keywords of different match types in the same ad group, or in different ones?
I knew about conversion tracking, but adding the conversion tracking code to our web pages was an adventure. The website was coded in HTML, so I had to add the conversion script to the page in Adobe Dreamweaver, email it to the support manager who would FTP it up to the website.
If you don’t know what any of that means, then neither did I at the start! Nothing was as simple as it seemed in the books. Everything took longer than expected.
Fully aware that our website needed significant work, I created separate landing pages (which I designed myself in Dreamweaver), offering a free CRM email course. People could opt-in to an email series exploring the benefits of CRM. I wrote the course myself in evenings and weekends, and set it up using the email system AWeber.
None of this work was a spectacular success. After reading Perry’s work, I was obsessed with generating ‘conversions’, which I struggled to do. Failing to understand how company budgets are set, some months I even failed to spend the full £1000, figuring (correctly) it was a waste of money.
Eventually the sales manager (who understood a bit about AdWords, and more about company budgets than I did), advised me to ‘just spaff it’.
But many of the ingredients for success were in place. More next time.
I’ve been thinking of starting my own ‘Google Translate’ service. Not a language transcription service… but rather a simple way to understand everything Google tells you.
For example (some of these are directly from the Google Ads blog)…
Google says: “You should raise bids on mobile”
Rob says: “Unless you have data to prove otherwise, you should probably lower bids on mobile”
Google says: “We’re enabling a safe digital advertising ecosystem”
Rob says: “We’ll repeatedly disable your ads for no reason”
Google says: “You should optimise ads for clicks”
Rob says: “You should rotate ads indefinitely and optimise for profit”
Google says: “Build a better mobile experience”
Rob says: “Create a fast-loading mobile landing page, you dummy”
Google says: “Disney and Google expand strategic relationship”
Rob says “(Not even gonna comment on that one – breaks translator)”
Google says: “How evolving user patterns drive new ad experiences on YouTube”
Rob says: “We’re gonna change the names of all the YouTube ads formats every seven seconds”
Google says: “An insider’s look at the latest Google Ads innovations”
Rob says: “Watch out! Here’s what we’re about to change in the interface next!”
More translations available on request.
One way to improve your ad performance on any network is to focus on a smaller geographic region.
Targeting the whole of your country (assuming you sell nationally), or even multiple countries is a fine way to test an idea. But if you’re doing that long term you’re probably missing a trick.
I was thinking about this the other evening, watching a quiz show called The Chase…
The format of the show is that four contestants answer questions individually in a cash builder round, where each correct question is worth £1,000 for example.
They then go up against a professional quizzer, and have to answer a number of multiple choice questions correctly to outrun the Chaser, and proceed to the final chase. If they’re caught by the Chaser, they’re out.
They can choose to play for the amount they earned in the cash builder round. Or they can go high or low. High moves a step closer to the Chaser, and earns them about ten times as much money if they get through. Low moves them a step closer to home, but decreases the money they take back to the team.
Before deciding which offer to accept, they turn around and ask the other team members what they should do…
In the British show, the suggestions are nearly always conservative. You’ll hear things like “you’re a good player, we want you back, just take the low offer.”
In the American version of the show, you’ll hear things like “YEAH GO HIGH MAN! You’ve TOTALLY got this…”
In the American show the Chaser will sometime offer a ‘super offer’, where the contestant can move a step closer to the Chaser still (for mega big bucks).
How does this relate to your ads?
RUNNING THE SAME ADS TO THESE TWO GROUPS OF PEOPLE WOULD NOT WORK.
Huge cultural differences exist within your country – not just internationally. This is often an easy if laborious way to improve ad performance.
And a reason you might want to have multiple copywriters working on your ads.
P.S. Every time someone corporate tells me they’re responsible for the “EMEA region” (Europe, Middle-East and Asia) I puke in my mouth slightly.
P.P.S. Europe is not a thing (see Brexit).
If you’re running remarketing ads across multiple networks, it’s critically important to document your ads.
I wanted to share a quick video today of how I document mine. If you do this as you go, it only takes a few minutes. And it saves you hours of head-scratching later on, especially if you need to pass your ads management to somebody else.
The document template can be found here. Select file and ‘make a copy’ to save it.
Have you ever created an experimental Google ad, only to forget about it until the monthly invoice arrives?
I have. It’s more common than you might think.
Here are three tips to stop this happening…
1. When initially testing ideas, use the start and end date feature
Each of your ads campaigns should be in one of three states: experiment, optimise or scale.
The experiment phase comes first. You’ve had an idea, you don’t know if it will work, and you need some data.
At the optimise phase you have that data, and are working to lower cost per conversion.
At the expansion phase you’ve achieved a satisfactory cost per conversion, and are ramping things up.
All experimental campaigns should use start and end dates, usually of not more than 7 days in advance.
If you’re testing many ideas, that provides a fall back in case you forget to switch off an unsuccessful test. (It sounds silly, but it’s exceptionally easy to do). Otherwise you only realise when the invoice comes at the end of the month.
This principle applies to Facebook ads also, at the ad set level. And also to LinkedIn and Twitter ads.
2. Make use of automated rules
Once you have a fairly regular number of conversions, you can set automated rules to pause an ad group or keyword if performance drops below a certain level.
You can also choose to have an email sent to you, which is usually my preferred option. (Often I’ll want to diagnose a keyword before pausing it).
3. Make use of custom alerts in Google Analytics
If you look under the ‘customization’ menu, you’ll see an option called custom alerts. Create alerts to notify you if bounce rate goes above a certain threshold for your paid search visitors. Again, this is simply an early red light warning to alert you about a possible website problem.
These three tips are free, and take a matter of minutes to setup.
Do you have any more?
A client forwarded an email to me on Monday. Someone he knows had paid £7,500 to a Google Ads agency… and was horrified to later discover they had only spent £3,700 of this on ads.
If you’re thinking of hiring help with your Google Ads (or any ads in fact), I have a few rules of thumb:
1. Always pay for your own ads
Never pay a chunk of money to an agency to run ads on your behalf. By doing this you lose visibility into what is going on, and you lose control of your own data. Any agency you work with should be able to send you something called a ‘client manager invite’, which grants them access to your Google Ads account for as long as you wish to give them access. You then pay the ads bill directly, cutting out the middleman. You should pay an agency for their time, support and expertise, not for ad spend.
2. Educate yourself
In my experience, most agencies lack introspection about which parts of the Google Ads machine to apply under different circumstances. Most under-use remarketing. Almost all test an insufficient range of ad creative. Whoever you hire, you’ll get better results if you yourself have a good understanding of these things.
3. Fee structure
My preferred ways to bill a client are either flat monthly retainers, or a commission arrangement is specific conversion actions are measurable. Or a hybrid approach of the two. As with these things, there are upsides and downsides to these approaches. The retainer approach is simplest. The commission approach is fairest.
As a general observation I’ve found clients tend to resist the commission arrangement. Partly due to complexity, and partly a reluctance to share the spoils. Which frankly baffles me.
Many agencies still charge based on a percentage of ad spend, on the basis that higher spend takes more time internally to manage. This is only true under certain circumstances (e.g. if you’re incompetent), and is entirely dependent on click prices and conversion rates.
If your click prices are £15 per click instead of £2, why should you pay more for someone to manage it? This arrangement also incentivises the agency to spend more regardless of results. I’ve seen agencies max out spending on brand name keywords, which really should be excluded from the billing arrangement.
4. Consultants vs agency
It’s worth considering that no one individual can specialise in the entire Google Ads machine (Google Search, Display, Shopping, YouTube…)
I myself specialise in Google Search, remarketing, ad writing, and customer nurture. I’ve dabbled in Google Display and YouTube, but I’m not an expert. Same for Shopping ads. I’m good at text ads, but average at image ads. My video creation skills are ropey to say the least.
Before you hire someone, ask which parts of Google Ads they specialise in. If they say ‘all of it’, then you’re probably wise to walk the other way.
5. Jump into the saddle when you need to
If you completely abdicate responsibility for your ads, you’ll almost certainly leave money on the table. From time to time, don’t be afraid to:
Scrutinise your conversion numbers
I’ll be talking more about these things at next month’s Pie, Peas and Google Ads training (Sheffield, 8-10 May). There’s still a few places if you can make it. Positive ROI on your training fee guaranteed.
Do you spend money on Google, Facebook, LinkedIn or Twitter ads? If so you should consider the following tactics…
Warning – this is advanced level stuff. Probably not for beginners.
1. Upload your customer list to Google, Facebook, LinkedIn and Twitter.
If you have a large customer list, you might want to only upload your highest value customers. To do that, I would export your customer file from your CRM or accounting system, and add a column to your spreadsheet for a ‘RFM score’ (recency, frequency, money). This is a weighted score that ranks customers in your list by recency of purchase, then frequency of purchase, and then total spent. Total spent is actually least important as an indicator of likely future purchases.
Sort your customers by RFM score, and upload only the top 20%. This limits your risk by focusing on your best customers, not your entire customer database. Your bottom 20% probably lose you money, so you don’t want to target them.
2. Run upsell ads to your customers.
You could offer time-sensitive bonuses in ads to existing customers, or incentives for making referrals. This is the highest ROI advertising you will ever do.
3. Upload a separate prospect list.
If you have a prospect list (in your email marketing or CRM system), upload them to a separate audience. Test running loss-leader type ads to this group, or free trial offers. Anything that will get them to make the first purchase.
4. Create lookalike audiences of your best customers on Facebook and Google.
On Facebook create a 1% lookalike audience of your customer list in the countries you sell in. Facebook will then build an audience of 1% of the total population that look like your uploaded list. (Which is why it’s best to start by uploading your best customers in step 1, not all customers).
Google will automatically create a ‘similar audience’, which in principle is the same thing.
You can’t currently create lookalike audiences on LinkedIn or Twitter. (Boooo).
5. Use your lookalike / similar audiences as a layering tool.
For example, if someone searches for one of your keywords AND is in a similar audience to your best customers, double your keyword bid.
On Facebook, try targeting people with relevant interests who ALSO look like your best customers.
Essentially you are giving Google and Facebook valuable guidance on who specifically you are after. Most of your cold ad campaigns (targeting people who do not know you) can usually be layered with a lookalike or similar audience.
6. Use custom affinity audiences on Google.
If you go into the audience manager in Google Ads, and select the ‘custom audiences’ tab, Google will usually have created something called a ‘custom affinity’ audience based on your web traffic. In mine, they have identified ‘CRM, marketing & business’. Which I’m actually quite impressed by. (Shockingly, Google seems to understand me better than I do!)
If the audience looks relevant to your business, try using this as a layering tool also (can be used with Google Search, Google Display and YouTube ads). For example, maybe target websites on Google display about specific topics, where the visitor is also interested in CRM, marketing & business (or whatever your custom affinity audience is).
7. Define your own custom affinity audiences.
In the screenshot above, click the blue plus icon to create your own custom affinity audience. In the screenshot below, I’ve manually added the interests ‘spirituality, literature, storytelling and marketing’. Google will then build you an audience of people with these interests.
You can try targeting these audiences directly on the Google Display network, but mostly I would layer them with other targeting criteria.
This is seriously advanced level stuff that none of your competitors will be doing. Also by focusing on customer lists, lookalikes and custom affinity audiences, you’re running ads in a multi-dimensional way your competitors can never copy.
Got questions? Let me know.
I live in Sheffield, Yorkshire. People round here have a ‘say it how it is’ perspective on life. Things they might say about Google Ads include:
“Ten pounds a click? ‘Ow much? Thieving bastards…”
“It were proper expensive that pal…”
“It tha’s running Google Ads, tha’s bloody daft…”
“I phoned Google earlier and spoke to some reyt sillyarse…”
“How the bloody hell d’you make this work?”
“Keeping the missus in rags, this adwords thing…”
“There won’t be no holiday this year, after that Google bill…”
If you ever find yourself uttering the semblance of these things, you might like to know that I’m holding a three-day Google Ads workshop, in Sheffield in May.
The workshop is called Pie, Peas and Google Ads, and it’s 8th, 9th and 10th May. Two days of training, and a third day blocked off for implementation before you return to work. You’ll leave with a working Google Ads account, not a headache and reams of confused notes.
The promise is that if you show up and put in the work, you’ll more than cover your investment in the training, in ad savings or increased profitability. Otherwise I’ll work with you personally after the event until you do.
(This is the only Google Ads training you’ll find where the trainer has a real vested interest in your short-term success…)
Plus if you’re fast there’s a handful of early bird places available.
It’s worth a read even if you definitely can’t come.
Many marketing agencies offer a Google Ads audit to solicit business.
Often these audits are about seven pages long or more, pointing out every single ad extension you could be using but aren’t. (Stupid you).
I have a slightly different approach. In looking at a Google Ads account, I’ll ask:
This one change nearly always comes from looking at your most expensive keywords.
Order every active keyword in your account by total spend. Make sure that cost and impressions are shown in the table. Then order all your keywords by cost (click on the column heading).
Normally the keywords with the highest spend will have the most impressions, but you might want to check out keywords with a high volume of impressions too. High impressions often means a keyword has been left on broad match by mistake.
For high cost keywords, ask these questions in order:
Ignore all other keywords for now, and just start with your most expensive.
You won’t end up with a seven page audit by asking these questions, but you’ll have something actionable to work on straight away. Which frankly, is far more important.
If you’re spending money with Google and none of the above made sense, email me.
Grab a paperback copy of Maze Remarketing: The 80/20 Approach to Profitable Multi-Channel Retargeting Ads. Just 1 penny plus cost-price worldwide shipping.