Why AI based PPC automation is far superior to Rules Based Engine
Imagine you plan to drive across your country and visit some cool places along the way.
What is the most important technology you would need to have a successful enjoyable cross country trip?
Most people take this for granted, but it’s our GPS. we don’t even need a separate device for it, it’s built into our phones.
Twenty years ago people had to print out their maps, draw the route and and follow it. They had no idea of the traffic or the road closures or any alternative routes or diversions due to changing conditions.
So why I am telling you this?
Because people using rules based system are a lot like people downloading their maps and planning their journey rather than let an automated GPS show you the way.
In this blog post I am going to walk you over some ways that Rules based PPC system is destroying your PPC game.
- Death Spiral
Of all the worst things that could happen when you use a rule based system this one stands out as the worst. So what’s the death spiral?
Imagine that in the rule based system, you set your bid to lower when the ACOS is higher than target. Lowering the bid will then further lower your sales which in turn lowers your ACOS even further. This trends then spirals out of control until the bid for your keyword is so low that there is no sales. There is no recovering from this.
- Seasonality changes
Some products have seasonality sales. Like winter jackets or snow shoes.
How do you account for that with rules based system? One way would be keep a list of all the products and their seasonal sales pattern and then change the rules for each product during their off season and peak season.
Yeah no thank you. Too much work. I would rather the AI do it automatically and more precisely.
- Weekend/Weekday & Holiday changes
One more trending pattern. How do you accommodate rules for different bids on weekends as well as holidays?
The chances of “Death Spiral” we talked above increases drastically during holidays and weekends when people are clicking our and raising your ACOS.
Your manual rule-sets are going to instinctively decrease bids without considering that it’s a holiday or a weekend.
Unless…. You login every weekend and holiday and change the rules. I am pretty sure this is not what you want to do and ruin your holidays and weekends doing boring mundane tasks of ad management. Leave it to the AI
- Competitor changes
The most overlooked aspect for the AI vs rule based systems is that rule based systems do not consider competitor landscape. Competitors are ever changing & dynamic. Their bids ultimately affect your acos because ads are an auction based system. Having fixed rules means that you are fixed and rigid in a ever changing dynamic sea of ads auctions.
- Automated day parting
This is simply impossible for rules based system. The most popular way that almost all sellers achieve day parting is to setup a system where your campaigns are active from 9 am to 5pm for example or even different times on weekends.
This is a big mistake since different products have different buying times. Also prime time for shopping is 8 pm and you are missing out on the best sales by switching off your ad during prime time.
It could be possible that some of your products sell best at midnight on weekends and some products sell better at 5 am on Mondays.
The only way you could figure this out is if you went through the sales history of all of your products and analyzed the sales trend. And even if you got the sales trend there is simply no easy way to setup complex rules like this with rules based engines.
It would work when you have 1 or two products. But as you grow in size and products, its utterly impossible to have dynamic bids with automated day parting.
AI based system can handle this easily and boost and suppress bids for thousands of skus in seconds depending on peak and off peak time.
- Historical performance
How do you handle historical performance in rules based engines?
That was a trick question. The answer is, you don’t.
Rules based systems cannot handle the fact that your current poor performing keyword was a star performance one month ago. It doesn’t need as harsh treatment as other genuinely poor performing keywords.
Sometimes its just a rough patch for the keyword. Don’t take a sledge hammer rules based system and pummel it.
- Black hat manipulators and bullies
Another overlooked aspect in the dog eat dog world and secret among black hat PPC optimizers. How this works is Black Hat PPC agents temporarily raise their bids too high and start winning sales at a loss. This increases their conversion rate, click through rate and their sales rank. Then they gradually drop their bids but in the meantime your rules based engine is initially struggling with lack of sales and then suddenly an increase in ACOS. Once your rules based engines start dropping the bids without considering the historical performance of your keywords, the performance of your other products, the performance of your account in general it enters the death spiral.
- A Holistic approach.
Profitable ads is not just setting a target ACOS. Much more things needs to be considered. The particular keyword historical performance, your campaign performance, your whole account performance, the weekday/weekend, holiday, peak hour and off peak hour. AI systems have around 100 parameters to consider before making a change while rules based system just have a narrow focus on ACOS.
- Don’t flog a dead horse.
There comes a time when plenty of your products are doing exceptionally well but some of your products are just bleeding ad costs. Even after couple of seasons, the sales are low and ad costs are high. Take a look at the CVR and CTR that AI based systems are high lighting. Can you improve it?
If not, maybe its time to finally give up that dead product and find new winning products?
Sure, printing out maps and getting lost in the woods is fun just like the “good old days”. And we recommend you to do it when driving cross country.But running an ecommerce business is not something you should play around or risk it.
Don't focus on the good old days, look forward towards the good new days