Conversions – a function of clicks or impressions?

Projecting paid search performance or PPC conversions as a function of impressions is problematic because any performance estimation,  is ultimately a function of searcher behaviour.

Definition of an impression, taken from Google AdWords:

Each time someone searches on Google or the Google Network and your AdWords ad displays, it is counted as one “impression.” The “Impr” located on your reporting statistics shows the number of impressions for your ad.

However, when someone searches on Google using Google Instant, ad impressions are measured differently. When someone uses Google Instant, ad impressions are counted in these situations:

  • The user begins to type a query on Google and clicks anywhere on the page (a search result, an ad, a spell correction, a related search).
  • The user chooses a particular query by clicking the Search button, pressing Enter or selecting one of the predicted queries.
  • The user stops typing, and the results are displayed for a minimum of three seconds.

Does an impression mean that someone actually sees the ad?

The answer here is no — just because the ad shows up on the search engine results page, does not mean that it has registered with the searcher.

If someone is using Google Instant and typing in queries, an impression may be counted when the searcher pauses for 3 seconds and then continues typing.

They may then have just glanced at the page and continued to search, potentially changing the search query to something not relevant to the product offering.

Three steps away from action

An impression is three steps away from action – first someone has to register your ad, look at it, decide to click on it and then interact with the landing page.

Without clicking and getting to the landing page, there is 0% chance of a conversion, because the person has to be physically on the landing page in order to take the desired action.

A percentage is only a number, so let us put a face on statistics

Conversion rate, is defined by conversions / clicks
Having a 1% conversion rate, means that for every hundred people that click on the ad, 1 will continue through and convert.

CTR (click through rate) is the ratio of impressions to clicks, defined by clicks / impressions
Having a 3% CTR means that for every hundred people who saw the ad, 3 will click on it

The above estimates hinge on a 100% confidence in the 1% or 3% estimate of a group of 100 people. (in a linear timeframe)

Statistically, that confidence is much lower and never 100%, so any percentage estimates will always vary, because you cannot predict human behaviour.

Using impressions to calculate conversions makes three assumptions:

1.       Searcher will see the ad
2.       Searcher will click on the ad
3.       Searcher will convert on the landing page and take desired action

This is a statistically flawed argument because the confidence that someone will take all three actions is much lower than the assumed 1% conversion rate for someone who does click on the ad.

In fact, using the above averages, there is a 0.03% chance that someone who has not clicked the ad, will convert. Thus, the projections will be unrealistically inflated with no statistical probability that it is even possible to reach the projected conversions when calculated as a function of impressions.

  • Share/Bookmark

PPC – It’s Relative

It has been a while since I have tweeted, networked on LinkedIn, made updates on Facebook or, even, written a blog post. It is this last point that perturbs me most, because as I have said before, blogging every three months more closely resembles a newsletter, than it does a blog. I should either eat my hat or start writing.

No Absolutes in PPC

Sometimes I am asked questions that go something like this:

-       If we double our monthly PPC budget what percentage impact will that have on conversions?

-       With our current PPC spend what should our performance be and what should our conversion rate be?

Though paid search is heralded for its measurability, one element it does not share with traditional media is its predictability.

Let’s take for example running a full-page ad in a monthly magazine.

The cost is fixed; the reach (number of people reached by single is issue), circulation (number of distributed copies of a magazine) and coverage (percentage of a population group reached by a magazine) are just a handful of magazine metrics that are widely known and shared by the magazines themselves.

Based on historic performance or industry averages, your company will also have a good idea of the return to expect from the magazine advertisement. There are few variable metrics – if you increase budget, ceteris paribus, you can calculate the expected return on investment.

Unpredictable PPC

When launching a new PPC campaign, where little historic data is available it is more difficult to predict the impact of budget increases or provide a forecast of performance.

There are more variable factors in PPC that are in your control, but that also makes the performance outcome more variable. The paid search bidding model also changes the landscape because sudden moves by competitors can also disrupt forecast predictions.

When setting goals and internal benchmarks for your paid search strategy it is important to set your goals based on the relative past performance of your account. Benchmarking against best-in-class numbers may be out of reach for your PPC account in the short term and focusing on absolute numbers will distract from the positive gains you are making.

If you are posting a 1% conversion rate, it may be unrealistic to aim for 5% in the short term (because you read somewhere that this is a good number towards which to aim). Instead, what you should be looking for is a positive trend, this may be slight month-over-month (MoM) but if you are focusing on driving performance then year-over-year (YoY) metrics should show a positive trending improvement.

Your best measure of success in your PPC account is your own YoY data rather than focusing on the 800-pound gorilla in your industry that may have 10 times your budget, can outbid you and dominates the top positions for your keyword basket.

The point is, there is no magic formula to determine how much return a specific budget should yield. It is about how much the budget could yield and this in itself is an iterative process. In PPC, it’s not about good or bad, it’s about better or worse. So if you moved your conversion rate from 0.8% to 1%, you are doing better, which is good. :-)

If you are looking for benchmarks and are willing to spend a little bit of money, start with some quality reports from MarketingSherpa.

  • Share/Bookmark

My PPC Ad Position Has Moved, Again!

Angry SEM ClientThere’s a great deal of advice out there about selecting the right SEM agency so that your business reaps a positive return from the online marketing, but what about the SEM agency getting the perfect client?

As difficult as it can be for clients to work with agencies, there is a plethora of clients that give agencies significant headaches with continuous phone calls, repetitive questions and unrealistic expectations that impact the success of the project. Are you expecting your SEM agency to deliver PPC miracles on an unrealistic budget? Have you set SEO ranking expectations on an unrealistic timeline?

If you are one of those clients who constantly check their PPC sponsored listings, you should be aware of some of the factors that may affect your PPC ad positioning. So, before you pick up the telephone to scream at your SEM strategist because you typed in some important keywords but only the competitors’ ads come up, consider the dynamics that may affect your PPC ad position.

Performance of Your Ad, Relative to Search Query
The relative performance of the potential PPC ads that could be shown for each search query are dynamically evaluated by the search engines. What this means is that the performance of your ad will be measured against:

  • CTR of your ad relative to the specific search query – what is your historic click-through-rate for that ad variation?
  • The performance of competitive ads relative to yours for a search query – how well have your competitors’ ads performed relative to yours?
  • CTR performance for each keyword in your ad groups – what is your historic click-through-rate for individual keywords?

What you need to bear in mind is that each search query is dynamically evaluated, which means that if you type in a keyword for your business now and check the results, in all likelihood when you check that same keyword query in 5 minutes the ad positioning will be different. A number of elements interplay dynamically to determine the ad rank of each competitive ad vying for a ranking on a particular keyword.

Maximum Bid
Maximum bid is the single most influential factor under the advertiser’s control that drives ad positioning. If you would like your ad to be at the top of sponsored listings in the number 1 position, setting a high enough maximum bid on keywords can ensure this.

Why? Maximum bid and Quality Score are the elements of Ad Rank, i.e.:

Maximum Bid x Quality Score = Ad Rank

Of course ranking in the number 1 ad position just for the sake of being number 1 cannot even be considered a bidding strategy. Your ad positioning should be driven by business metrics such as desired maximum customer acquisition (CPA) and it should be within the limit of your maximum CPC. Simply bidding up for higher rank positions may very quickly render the PPC efforts unprofitable for your business.

So, if you are one of those clients that obsessively like to see their PPC ads in the 1st position on the SERPs, stay calm if you see your ad in the 3rd of 4th positions. This may be the optimal position for your ad based on driving the highest level of relevant traffic to your site, within budget and at an acceptable cost & conversion rate within your business model.

Geographic Targeting
Be mindful, that if a specific campaign is targeting a geographic region that lies beyond the physical location from which you are checking your ad positioning, it is a good thing if you do not see your ads appearing!

Specific geo-targeted campaigns allow for tighter location-based targeting of your market. For this type of PPC campaign to be successful, campaign settings need to be set up so as to minimise “leakage” of ads appearing in non-relevant geographic regions.

Dayparting by You & Your Competitors
Avoid embarrassment – if your target market is in the B2B buying space and your customers are known to be the most active during business hours, then give your search marketer credit if they are rotating ads during specific hours, instead of ‘freaking out’ because the ads are not showing at 7:00am.

Also, bear in mind that the dayparting bid strategies of your competitors will affect your positioning. If a few competitors, with higher bids start appearing at a set time while your ad is showing, it may serve to drop your ad rank positioning. The greater the number of advertisers competing for a keyword, the more variable ad positioning will be.

A/B Testing of Ads
If you are testing your ad variations, which you should be, invariably there will be a weaker ad and a stronger ad. The ad with the higher CTR and greater perceived relevance may then have a higher ad position than the other one. When you type in a search query into the search engine, you cannot control which ad will be shown for that query.

Nebulous Search Engines
The fact is that the search engines have their unique algorithms according to which they may practice random testing, ad rotation, algorithm adjustments and Quality Score allocations. Not everything affecting ad position fluctuation can be explained according to a checklist of possible causes. Sometimes, the cause of a rank drop is due to the cumulative effect of a number of elements, some of which the advertiser can influence and others over which the search engines maintain exclusive control.

So, before you pick up that phone, consider that your SEM agency’s time would be better spent managing these dynamic elements of your PPC campaign, rather than having to explain the fundamentals to you, again!

  • Share/Bookmark