Introduction.

‘New normal’, ‘roadmaps’, ‘adaptation’, ‘unprecedented change’ – these are all terms that have become commonplace in everyday conversations over the past year during the COVID-19 pandemic. As we’ve changed our behaviours, whether that’s panic buying rice and toilet roll in the supermarket, or choosing a desk for your newly converted spare bedroom/office, our purchase cycles have become different too. Now the roadmap out of the pandemic and lockdowns have been drawn up, businesses are yet again having to create plans that could change at short notice, with various scenarios shooting off in different directions depending on potential timelines and budgets.

Pay-per-click.

Pre-pandemic there had been a clear focus and progression towards automation. With consumer behaviour more predictable than ever before, we’ve had a wealth of data to inform machine-learning of what was likely coming next, and how we wanted the machine to react to this. Paid Advertising and specifically PPC advertising was following this path, moving towards automated techniques wherever possible. Bidding strategies, keyword targeting, and ad copy are all prime examples where Google has been keen to move the industry towards a more automated approach, to save more time for more broad, high-level strategy discussions.

Despite this, the COVID-19 pandemic has shown that the human element of advertising is still as essential as ever, we’ve seen plans change at the last minute at Christmas, demand spiking for certain products, and budgets needing to be adjusted at short notice. Automation will adapt once it sees a demand change, but the uncomfortable truth is that it is not built for these unpredictable situations. The very nature of automation is that it will indiscriminately follow a route it has been set to, the more uncertain and challenging that route is, the harder it will find to navigate to its destination.

This past year has been one of the most uncertain for decades, and human challenges are inevitably ones that automation will have to deal with too. Automation is not aware of COVID-19, and unlike a human, it cannot dial into a business meeting or listen to a press conference to hear what impact it is having on business decisions. By the time you have set up a framework to help guide it through, you likely could have tweaked a manual structure in less time, as manual campaign structures offer more control and flexibility to navigate around these challenges.

This article will explain the pitfalls of PPC automation in a present and post-COVID world, how you should navigate them, and what your team or agency should be doing to help achieve your business goals.

Learning Phases

A learning phase is a period where the algorithm is calculating how best to reach the target that has been set. Whilst this is usually associated with the start of a campaign, it can also be activated if a ‘significant edit’ has been made to the campaign(s). At a time where change is more likely than usual, the risk of kicking in the learning phase is heightened.

Whilst performance won’t decrease significantly as a result of the learning phase, it’s unlikely that the target you set (e.g. a 5 ROAS/£2.50 CPA) will be hit during this period. Indeed Google even notes in their official documentation that ‘you may not want to measure performance until the learning period is over’ https://support.google.com/google-ads/answer/6263057?hl=en-GB.

In a standard setup with manual CPC (or enhanced CPC), quick changes can be made with far more flexibility. For example, if you need to reduce your budget by 20% with immediate effect, reducing your CPC’s by this amount is an effective change, with the intended result being actioned immediately. Whilst in an automated setup you would be able to change your target ROAS or CPA by 20%, the algorithm will inevitably need more time to re-route and produce the desired result.

Automation: Rigid when it comes to sudden changes, with short-term reset needed to achieve the desired target.

Manual: With the right experience, manual campaigns can be very quickly changed to achieve a new target, with lots of different levers (bids, device adjustments, audiences) that can be pulled to achieve change almost immediately.

Desktop vs. Mobile

Mobile-first has justifiably been the focus in the past few years pre-COVID, as technology continues to improve and people have become more and more transient in their routine. However, with lockdowns keeping potential customers in their homes with more time, there has been a reset to the ‘browse on mobile, convert on desktop’ mindset.

Whilst automation will over time see this change and bid more aggressively on desktop, it is not something it can understand immediately, especially with well-established campaigns that have previously been mobile-dominant. Using a manual approach, this is something that can be taken into consideration immediately, with the option to apply a positive bid adjustment on desktop, or conversely a reduction on mobile or tablet.

At a time where restrictions are changing at different speeds, flexibility to make quick changes on segments such as devices is essential. A manual campaign allows for a proactive approach, if you are aware of the roadmap of restrictions easing or being tightened, changes can be made in advance, and planned. Automation is restricted to being reactive, once it has spotted a behaviour change it will adjust, but a competitor using a manual strategy would already be a step ahead.

Automation: Restricted to reactive changes based on behaviour once it has already started to take place.

Manual: With planning and research, you can align your bid adjustments ahead of time, monitoring when behaviours are likely to change, and actioning proactively.

Demographic/Geographical changes

A similar theme to the above, but the change in geography has particularly impacted B2B pay-per-click advertising. With employees and c-suite decision-makers forced to work from home for extended periods, this has led to many working outside of metropolitan areas, particularly in the UK where commuters into London are now working from suburban or even rural areas.

Therefore, location performance will inevitably change too, with decision-makers suddenly far more spread out than ever before. With B2B typically producing exponentially higher CPC’s than B2C, it is vital to eliminate wastage, but also maintain visibility for potentially lucrative customers where volume is less. Again, unfortunately, automation’s reactive approach is incompatible with this change, with algorithms reasonably applying a premium on searches within metropolitan zones, and discounts on rural areas.

Automation: Slower to react to substantial changes in behaviour, leading to under-bidding on potentially lucrative situations.

Manual: By using a manual approach, you can change your targeting at short notice, reducing or increasing depending on changes in customer behaviour. By analyzing performance during previous lockdowns, you can adjust campaigns in line, ahead of time.

Interest/Intent-Based Searching

As the roadmap out from lockdown becomes clearer, we naturally start to plan, whether that’s a big purchase such as a foreign holiday, or just finding an appointment to get that overdue haircut!  Whether big or small, some of these decisions will be taken in advance, and some will be a panic-buy at short notice.

For those that are more likely to be planned, they will often involve multiple online touchpoints, through the typical browsing phase to ultimately completing a purchase. Whilst in a normal situation you would expect automation to be able to deal with this pattern of behaviour as purchase cycles would be relatively similar throughout the year, this is unfortunately not a normal situation.

As seen in the UK, provisional dates have been pencilled in months in advance, for example when larger weddings or foreign holidays could happen again. As these dates move closer, interest becomes intent, and every news story or subsequent announcement can lead to a change in the interest/intent balance. An account using advanced manual bidding techniques can flex up and down, reacting to announcements and news as soon as they happen, as opposed to automated strategies that will only react after conversions/sales begin to come through. As mentioned at the beginning of this article, an automated account cannot listen to a press conference and apply context to the announcements made, a human applying manual bidding techniques will be able to take this information, act upon it and improve performance before the automated strategy has seen this change in consumer behaviour and recalibrated.

Conclusion

This past year has been one of the most unpredictable for decades, and certainly the most uncertain since the e-commerce/.com boom. Whilst automation has been able to replicate manual processes, they have been built to see what has happened before, and not for what is going to happen next. Whilst using manual techniques may take more time to execute, a well-structured, flexible account is far better equipped to react to uncertainty, allowing for more proactivity.

Whilst automation may have been the easy answer over the past couple of years, it is worth asking whether it is still the best option for now, and moving forward as we emerge from over a year of ‘new normal’. Automation may be time-efficient, but with more uncertainty to come, it is unlikely to be performance efficient.