Strategic Principles in Paid Advertising

Rationed views, gas pedals, per person frequencies, when to trust platform reps

This article is part of my Entreprenerd: Marketing for Programmers book, which is currently available to read for free online.

Topical Advertising vs. Interruption Advertising

This distinction is about the state of mind of the person seeing the advert—regardless of how they were targeted.

The quintessential example of topical advertising is a Google Search ad which gets triggered based on certain search queries (e.g., “nearby gas station”). These ads reach people that know they have a specific problem and are in the market to solve it right now. The advert is fully on topic, so advertisers can expect high click-through rates, high conversion rates…and high competition from all the other businesses that have arrived at the same conclusion.

Interruption advertising, by contrast, is displayed when the recipient is focused on something unrelated to the subject of the advertisement. They are watching YouTube, reading an online newspaper, or dotingly browsing the Facebook profile of a secret crush. These adverts shoehorn their way into other experiences and demand your attention. Examples include ads between videos, flashy banner ads, or sponsored stories showing up in your Facebook feed. Interruption advertisements are the online marketing equivalent of cold-calling; the marketers deploying them have a difficult job, as can be seen by the resulting click-through rates that are often an order of magnitude lower than those achieved with topical advertisements.

I would recommend you pick topical advertising for your initial campaigns because it’s conceptually simpler and usually more profitable (despite the high competition). But do not take this recommendation to mean that topical advertising is unequivocally better. There are cases when you’re better off with interruption advertising. One is where your business sells a new product that is unknown to your market; because there’s no way that anyone could be expected to specifically seek you out, you have no choice but to rely on interruption advertising, at least to begin with.

Another use-case for interruption advertising is when you’ve saturated your reach with topical advertising and now wish to extend your marketing to a larger audience only accessible through interruption advertising. Just because I don’t explicitly google “Radiohead concert tickets” regularly doesn’t mean that I wouldn’t buy them if presented an opportunity. Interruption advertising ensures that this opportunity comes my way.

Rationed Views

Today there are over 7 billion people in the world. But unless you’re selling water, only a small proportion of these people will have any interest in whatever it is you sell. This tiny group of potential customers is your intended audience, and it’s these people that you target.

To a greater or lesser degree, your advertising only enables you to reach a narrow sliver of this targeted audience on any particular day. Imagine you are selling guides to pass law finals. Your targeted audience will be all students sitting their law finals at the end of this academic year. But only a fraction of these students will google “XYZ exam” on any given day, meaning that only a fraction of your audience will see the adverts you configured to appear for those keywords. Despite your targeting, your access to the underlying intended audience is rationed.

By contrast, suppose you had a parallel Facebook campaign that displayed adverts to Facebook users who identified themselves as studying law this year. Because a much larger percentage of law students log in to Facebook every day than search on Google for a particular law exam, you will reach a much wider slice of the underlying audience on any particular day. Through this targeting mechanism, your access is rationed to a much lesser degree. (Of course, you’re unlikely to reach everyone in your targeted audience because some of them might not log in to Facebook at all while your campaigns are running.)

As with any attempt at binary categorisation, the full picture is messy; no targeting mechanism can be said to be fully rationed or completely unrationed. Day to day, membership of the Facebook law student group is mostly stable, with little in the way of migration in or out. If you were advertising to this group, nearly every member of the underlying audience would see your adverts after a week or so. By the end of that academic year, this audience would be totally saturated with your constant advertising…until, suddenly, the class disbands and gets renewed with the next batch of students, meaning that your advertisements become fresh again, at least for a week. As such, there are really two different rates of rationing, depending on whether you set your temporal resolution to days or years.

Despite its imperfections, the idea of rationing is a powerful one, and worth keeping in mind when thinking about your online advertising campaigns. Here are a few practical consequences and applications:

1. Possibility of gas pedals

Advertising to an unrationed group means that you can get through to the entire underlying audience in a short amount of time, so long as you’re willing to bump up your budget sufficiently. Putting your foot firmly down on the gas pedal like this is great for launching a product or triggering a rapid-action bulge in sales. But, once everyone in your targeted audience has seen your adverts 10 times, your sales will peter out and your advertising campaign will cease to be effective. At this point, the revenue potential of this audience has been exhausted, at least until the underlying group refreshes its members. As an advertiser, you should anticipate this dip and be prepared to put a brake on your budget as soon as you notice the party ending, as indicated by plummeting click-through or conversion rates. In short, this is no fire-and-forget advertising strategy.

Contrast this strategy with the one applied to rationed groups. The advertiser here must accept that there is no gas pedal they can floor. Indeed, throttling their budget would be disastrous. In allocating a massive daily budget to their campaign, all they will succeed in doing is bombarding their thin daily ration of audience with the same ad hundreds of times. Whatever advertising budget isn’t squandered in this way will remain unspent, sitting on the platform waiting for the next member of the target group to show up and be delivered the same advert for 110^th^ time.

Advertising to a rationed group has advantages though, accruing to the patient advertiser. They will be rewarded with a yearlong trickle of cost-effective traffic. Because they can assume that the people experiencing advertisements on any given day are seeing them for the first time, they needn’t worry about their adverts petering out due to saturation. Their campaigns will be low maintenance, and possible to leave unsupervised for long periods of time.

2. Targeting migration into the underlying group

The ideal targeting option focuses on recent migrants into your target group. The advantage is that these are the people that don’t yet own your products or haven’t yet seen your adverts. You’ll be able to reduce advertising waste in proportion to your ability to specifically reach these incomers through specialised targeting.

Let’s revisit our running example. If you had targeted all law students (from first year to finalist) instead of just finalists, then there would only be a partial refreshing of the underlying group each year, corresponding to the entrance of new first year students and exit of the finalists. But the old first year students would have become this year’s second year students, and the old seconds this year’s thirds, causing incomplete group renewal.

3. The need for anti-advertising-blindness measures

The human psyche is adept at tuning out irrelevant noise—literally, in the case of a rattly old air-conditioner whose din fades into the background noise; and figuratively, as happens when an advertisement ceases registering in our conscious awareness. This last phenomenon is known as “advertising blindness”, and marketers who don’t protect themselves from its effects will underperform compared to those who do.

Advertising blindness is more likely to turn up in unrationed advertising campaigns than their rationed brethren. When you squash so much advertisement into so little time, you increase the probability that the audience will be overexposed to the ads and will develop a perceptual blind spot towards them.

My own practical experience confirms the existence of advertising blindness. I noticed, when advertising to a relatively unrationed group on Facebook, that my campaign’s click-through rates started to falter only three days after the campaign began. Luckily, this downward trend isn’t inevitable—it can be reversed by introducing a few counter-measures.

The simplest of these is to create waves of advert variations. Once advertising blindness sets in, you disable the old adverts and swap them for the unseen variations. Since these new variants contain novel images and copy, they will be noticed afresh by your audience.

If you exit a room with a loud air-conditioner and return later that evening, you’ll become aware of its noisy din once again. This is because your selective blindness resets after the stimulus is removed. This suggests that advertising blindness can also be lessened by turning adverts on or off, or by relying on frequency capping, the subject of our next section.

4. Frequency capping

Marketing lore has it that an ad must be seen seven times before it is fully effective. There are remarkably few studies done to confirm that seven is indeed the magic number, so take this as a matter of conjecture rather than fact. For example, Herbert E. Krugam found that maximum advertising response was reached after just three viewings,[^7-6-1] whereas Thomas Smith believes 20 are needed:[^7-6-2]

The first time people look at any given ad, they don’t even see it. The second time, they don’t notice it. The third time, they are aware that it is there. The fourth time, they have a fleeting sense that they’ve seen it somewhere before. The fifth time, they actually read the ad. The sixth time, they thumb their nose at it. The seventh time, they start to get a little irritated with it. The eighth time, they start to think, “Here’s that confounded ad again.” The ninth time, they start to wonder if they’re missing out on something. The tenth time, they ask their friends and neighbors if they’ve tried it. The eleventh time, they wonder how the company is paying for all these ads. The twelfth time, they start to think that it must be a good product. The thirteenth time, they start to feel the product has value. The fourteenth time, they start to remember wanting a product exactly like this for a long time. The fifteenth time, they start to yearn for it because they can’t afford to buy it. The sixteenth time, they accept the fact that they will buy it sometime in the future. The seventeenth time, they make a note to buy the product. The eighteenth time, they curse their poverty for not allowing them to buy this terrific product. The nineteenth time, they count their money very carefully. The twentieth time prospects see the ad, they buy what is offering.

The ideal number of times an advert ought to be shown to a potential customer is more than once but certainly not “as many times as possible”. There comes a point when you are just throwing money away. Moreover, aggressively frequent advertising irritates people, in much the same way as an overplayed TV rerun or Last Christmas on the radio.

Frequency capping, a feature available on some (but not all) platforms, limits how many times a given person will be shown your advert on any given day. Once that limit is reached, this person won’t see your ads again, at least until tomorrow.

Frequency capping has an additional benefit beyond the ones mentioned already: It changes the distribution of your advertising reach. All things being equal, setting a low frequency cap (e.g., one impression per day) skews your budget towards showing your ads to as many different people as possible. Imagine the total size of your underlying target audience is 1,000,000 people but you only have the budget for 250,000 impressions over the lifetime of your advertising campaign. If you frequency cap at 1 impression per person, then your advertising platform can’t rush through your budget by showing the same advert a thousand times to whoever happens to be online. Instead, the platform must wait until someone who hasn’t yet seen your advert appears before it can display your advert and earn a commission.

When capping frequencies, keep in mind the entire experience of the person seeing the adverts. If you cap the frequency of one of your campaigns, this will not lessen the number of impressions from any parallel campaigns, either in the same or in other advertising platforms. Think about how your audiences do or don’t overlap. Furthermore, even the lowest possible frequency cap, one impression per day, will oversaturate its audience if the campaign is left running for months on end.

While I am glad that frequency capping exists, I think advertising platforms can do a lot better. My beef is that they assume that the best time to show someone an advert again is…the next day! I beg to differ. Surely the ideal frequency would be the same as the one trained mnemonists use for cementing facts into their long-term memories? If I learn a new word in French today, then memory experts advise me to revisit it tomorrow, next week, a month later, six month’s later, and then two years after that. Why can’t our advertising platforms follow the same tried-and-tested schedule?

The Progressively Rising Cost of Additional Leads

You are a retailer of pianos, the only one in your city. Today, 10 locals have taken the day off work and promised themselves to have ordered a piano by the day’s end. On top of this, a further 100 people are considering buying one, but haven’t made up their minds yet. They are still sitting on the fence.

The people who are determined to buy a piano today are likely to google with phrases that communicate immediate buying intention: “buy piano in Oxford”, “price comparison Yamaha Clavinova”. But those who are less decided will show less urgent and less direct interest through their search queries: “how to choose a piano”, “which brand of piano is best for a beginner”.

Advertising to those who want to buy today is a better use of your time and budget. These people don’t need convincing—they just need a price and a confident promise of next-day delivery. All in all, it’s an easier sales proposition. But there’s not so many of them.

Once you have exhausted this initial supply of eager customers, revenue stops growing. Now it’s up to you whether you want to branch out and advertise to the larger, second group of people who are vaguely considering buying a piano. With a sophisticated marketing campaign consisting not just of extended advertising campaigns, but also of weapons-grade sales copy and incentives like one-day-only discounts (that cut into your profit), you may convince one or two people on the fence to buy today. But all this effort has a cost, and the final tab shows that the price of selling to this second group is higher. The low-hanging fruit of the first group now exhausted, you must dip further and further into your pockets to continue making sales.

If the goal of business were to maximise margins and seek only the most efficient sales, there would be no point in financially straining to reach the customers in the second group. But this isn’t the goal of business; the goal is to generate the most overall profit. As long as you continue to extract a healthy margin when taking into account the increased advertising costs, it’s worth paying up.

This idea isn’t always apparent at first sight because our emotions can lead us astray. To see what I mean, take a look at the results for these two advertising campaigns run on Google AdWords:

(Let’s assume we are selling a product with no variable costs, such as an information product or software.)

| Campaign | Advertising Spend | Revenue | Google’s Cut | |-----------|---------------|------------|--------------| | Definites | £1,000 | £3,000 | 33% | | Maybes | £1,500 | £2,000 | 75% |

Table: "Google's Cut" is the perctange of your revenue that it charges you for advertising services

In the calculation above, you’ll notice that advertising to the Definites is much cheaper. Google takes a 33% cut—a lot of money to be sure, but still a tolerable burden.

Now consider how you feel about the results of advertising to the Maybes. Your instinctive reflex might be to feel ripped off: “Why the hell should I give Google £1,500 for every £2,000 I earn? They’re making more than me! I’m turning that rip-off Maybes campaign right off…. Good riddance!!” True to your word, you switch off the Maybe campaign and feel happy that those greedy a**holes at Google are no longer taking a bigger cut than you are.

A month later, you notice that your month’s profits have sunk 20%. Dismayed, you analyse the situation once again, this time adding a “Profit” column to your spreadsheet.

| Campaign | Advertising Spend | Revenue | Google’s Cut | Profit | |-----------|---------------|-------------|--------------|---------------| | Definites | £1,000 | £3,000 | 33% | £2,000 | | Maybes | £1,500 | £2,000 | 75% | £500 | | Total | £2,500 | £5,000 | 50% | £2,500 |

Table: "Profit" is simplistically defined as "Revenue - Advertising Spend"… remember that we have no variable costs here.

As you can see, advertising to the Maybes generated a profit of £500.When this campaign was turned off, it wasn’t like these leads would magically keep on visiting and buying. Without this campaign, the Maybes source of traffic froze to a halt, along with any profits they generated. By focusing on our outrage at Google’s disproportionately large cut, we neglected our own profit figures. By letting our emotions take hold, we made a bad business decision. We’re like the fool who, when offered “1% of £1,000,000,000” or “100% of £10”, chose to take the 100%.

Let’s continue this story. Now at peace with Google taking a large cut, your trust in the power of online advertising soars, so you decide to target a third group, the Question Marks. These are an even more speculative group than the Maybes, but you have reason to believe it’s worth a shot.

After a month, the results are in, and they aren’t good…

| Campaign | Advertising Spend | Revenue | Google’s Cut | Profit | |----------------|-------------------|---------|--------------|----------| | Definites | £1,000 | £3,000 | 33% | £2000 | | Maybes | £1,500 | £2,000 | 75% | £500 | | Question Marks | £9,000 | £6,000 | 150% | (£3,000) | | Total | £11,500 | £11,000 | 105% | (£500) |

Despite a large increase in revenue, profits have nosedived. Your business LOST money last month! Advertising to the Question Marks was a horribly failed experiment, so you react by radically decreasing the bids to this group, brainstorming alternative advert creative ideas, and maybe even deleting the whole campaign and just being done with it.

I depicted the effects of advertising to the Question Marks by neatly reporting them in a separate row. But this sort of orderliness isn’t the reality for many advertisers. Through poorly implemented tracking technologies or lack of analysis know-how, these advertisers might not be able to distinctly group their customers into Definites, Maybes, and Question Marks in their performance readouts. All these advertisers have is an end-of-month summary of their Adwords performance (i.e., just the “Total” row from our table). And if that total is negative, as it would be this month, this advertiser may become an ex-advertiser, switching off Adwords for good.

This is a sad outcome, because if they had carved their audience into groups like we had, they would have seen that the Definites and Maybes generated a handy £2,500 profit per month—money they would not have otherwise earned. This shows us that decisions about advertising performance should not be made on a “this whole platform sucks” level. We should instead break up advertisements into specialised groups and reconfigure our campaigns to reinforce the successes and banish the failures.

Switch On All Bells and Whistles

Advertising platforms are always releasing special new features that change how your adverts will appear to leads. For instance, Google Adwords has what they call “ad extensions”^7-6-3. These let you do things like insert 4–5 additional links into your ad; insert buttons for downloading an app; display a picture of a phone which mobile users can click on to call your office; add reviews; or insert “callouts”, Google’s term for short snippets of copy highlighting your key selling points (e.g., “free delivery”, “100% money-back guarantee”).

Unfortunately, there’s no guarantee that your sitelinks will work after you enable them. Eligibility is dependent on your advertisement’s rank and quality score, meaning that you may have some work to do before partaking in their advantages.^7-6-4

Another way to stand out on Google Adwords is to choose an attention-grabbing ad format.^7-6-5 Instead of sticking with boring old text ads, you might try image, flash, gif, or video ads.

Twitter have their own parallel series of ad extensions via the same “Twitter Cards” we saw in the SEO section of this book. For example, the video card plants a video into a sponsored tweet, while the lead generation card inserts a call-to-action button that, when pressed, sends that lead’s Twitter handle and email address to your company.

Turn as many of these extensions on as you can, if only for the primitive reason that they’ll make your advert take up more space on the user’s screen, thus grabbing you a larger share of their attention. As well as this “bigger is better” effect, extensions also give you another chance to convince your leads by means of providing social proof, extra links to your related products or services, playable videos, or simplified conversion funnels.

Unintentionally Competing with Yourself

Due to the auctioned nature of online advertising, it’s possible for two of your adverts on the one platform to compete with each other and thereby drive up your advertising costs. The effect would be as if another business were bidding against you for that same audience.

By leveraging the organisational features of your online advertising platform, you should be able to avoid this problem. Ads within any “ad group” don’t compete with one another, but do compete with ads from other ad groups. It is therefore advisable to avoid simultaneously targeting the same underlying audience through different ad sets.

Competing against yourself can happen in frustratingly subtle ways. For example, in Google Adwords, two keyword variations that seem distinct at first glance may in fact be fierce competitors. An ad group that targets broad-match keyword “sandwich grill” would compete with a separate ad group targeting the (also broad-match) keyword “panini grill”, thanks to the synonym matching power of broad match.

Staff Working for Advertising Platforms Are Experts in Platform Features, Not in Your Profitability

(This is mostly a Google phenomenon, so what follows is Google-centric.)

Google isn’t exactly known for its customer service. There is no way to contact its employees, come good or come bad. At best, you’ll be invited to post your question to an official Google forum and pray that an empathetic onlooker happens upon your problem and advises you. But don’t count on it.

Google’s stronghold is impenetrable to mere mortals, except when those mere mortals splash out on buying Adwords. As soon as your budget reaches a certain threshold (approximately 500–1000 pounds/month), armies of keen Google representatives magically appear to offer you regular, free consultations on optimising your campaigns. Beginners will be walked through the most basic optimisations for quick wins, intermediate users will learn about recently released features that might nudge them ahead of their competition, and pros will be reminded to stop obsessing over the little things and refocus on getting the basics right.

I would do nothing but gush about my love for Google’s free Adwords consultancy sessions, but for the problem of misaligned incentives. Google and its employees are paid whenever you spend money on advertising. The more you spend, the better it is for them. But from your point of view, the less you spend on advertising (for the same amount of sales), the lower your expenses and the higher your profits. These incentives are at odds with one another.

Now, don’t get me wrong: Google seeks long-term win/win relationships with their clients, so they certainly desire that your advertising be profitable for you. But profitability is not an all-or-nothing matter: It’s really a negotiation between you and Google about how much commission is reasonable in exchange for bringing in a sale. During these consultations, Google employees often advised me to raise my bids (i.e., to pay them more) in exchange for dubious “benefits”, such as ranking above my competitors’ ads or building brand awareness with hitherto unreachable customers, the ones for whom my previous low bids were insufficient to reach. But these advertising “benefits” were not worth the price they encouraged me to pay. And not once did Google advise me to reduce my bids—even when that was the wisest course of action.

[^7-6-1]: Herbert E. Krugman. “Why Three Exposures May Be Enough.” Journal of Advertising Research 12, 6 (1972): 11-14

[^7-6-2]: Thomas Smith, Successful Advertising, 7th edn, 1885


More Articles: Click here for full archive

The Bug Slip

A bug hunting ritual that makes things a little easier


Janki Method — Using SRS to Improve Programming

Using spaced repetition systems to learn and retain technical knowledge.


WWW vs. Naked Domain?

When starting a new website, is it better to choose domain.com or www.domain.com?