The initial promise of digital, at the time when it is still called web, was simple, clear and unstoppable. The first adopters of marketing with online communications used it endlessly. According to them, marketing via the web enabled you to measure everything that is connected via the web including page views, products views, etc.
The use of digital marketing has become a science…
By utilizing this measurement, early adopters were able to control the ROI of marketing investments dedicated to a web based channel unlike other channels whose ROI might appear unclear (since, measuring ROI of TV and radio campaigns has made tremendous progress, but at the time this was not the case). While this data did not change marketing habits right away, an incredible thing happened: the argument presented by early adopters was proven and worked…
For years marketing was a controlled mechanical function that engaged the same channels, advertisers and customers in a fixed cycle. Messages were communicated in mass in undifferentiated ways to entire markets that may or may not reach the desired targets. Marketers discovered that by creating customized messages they were able to reach desired target audiences with relevant messages.
Data was slow and complex to collect at first leaving marketers to rely on their intuition. Today, overwhelmed by a huge mass of data, they are now able to gather meaningful statistics, systematically measure results and utilize real-time analysis on rapid changes occurring in the market place.
With digital, their marketing has become scientific, and we can actually measure everything, as indicated in the original promise of the web. Is that not a great dream? A beautiful promise? Not so sure.
… But can we wake up from this dream?
At Hub’Scan, a French start-up company, we believe you must wake up from this dream, even though it may be very enjoyable. For many, the data we collect through the use of data analysis platforms and algorithms is so powerful and makes us feel so smart, yet the great majority of analysis we discover is simply wrong. Why? Due to human error … even our own.
Explanation: marketing data that feeds our analytics platforms generates millions of data every second, as a result, the data generated is simply constrained.
Yes… tagging operations are important and have become common practice, but it is tedious and tags created can be poorly made. Tag Management Systems (TMS) that promised to overcome this problem have not solved the issue of poor tagging. The TMS applies its “process”, but its updates are too small (each evolution of a website must use the process followed by the TMS, otherwise it will continue to file the same tag on a page whose content has changed).
Two different reasons and a single result: the data in our analytics systems are wrong.
With the explosion in the number of pages within an e-commerce site, and the runaway inflation of the number of products offered, promotions and special offers on a website remain competitive even while using tagging errors.
If our measurement and analysis systems do not become more efficient, we will integrate data less efficiently. As a whole, we like consistency in an increasingly scientific, coordinated, industrial system. Thus, the false data, processed on a consistent basis, generally leads to erroneous conclusions. Less is more … it gives less.
Marketers have invested heavily in the quality and intelligence of analysis models. It’s time to do the same with their raw material: the data and its reliability.
How Hub’Scan works, Eric Dumain: “Briefly, Hub’Scan will allow the construction of comprehensive marketing plans, monitor the compliance of these plans and improve the operation of the tags. With the slightest change in trends a notification is sent the head of the line product, advertising campaign or the site manager. To go further Hub’Scan will monitor the customer experiences and conversion tunnels to optimize the knowledge of visitor behavior. Best of all, like all Hub’Scan analysis (speed, http status, tags, TMS, data, videos, frames, forms, secure space …) it is used it to make a full SEO audit and say what, how, and where to optimize ! … Everything will happily feed your data layer and ensure that data is consistent with what is expected in near real-time.”
And the data proves it!
Your website has a promotional page. It is one of the most visited, because consumers like bargains. This month you are planning to launch a promotion of a fashionable smartphone. To analyze the performance of this promotion, you will mark the page in question with a tag indicating the product “Smartphone”. The content of the page (photo, price, blurb) will be associated with that tag. Whenever a tag is searched, the system will recognize the search and capture data indicating the content of the page has been viewed by a potential customer.
Then, once the promotion “smartphone” is complete, you replace the content of the page promotion describing the content of the new product on promotion. For example, a “coffee maker”.
However, you left the tag “smartphone” imbedded in the page content, which is a common mistake. Your analytics tools will therefore continue to believe that the content “smartphone” is displayed, creating data that indicates you are still promoting the old product to a potential buyer, each time the page is accessed. While the page you are promoting is now a coffee maker it will appear that coffee makers sales are soaring due to its presence in an inaccurate target market space.
At the same time, the smartphone you continue to sell on your site despite the end of the promotion, has returned to a normal life cycle: normal price (more expensive than when it was on sale) and original product page (less emphasized and fewer views than the promotional space). It therefore draws less traffic than when it was sold on promotion.
So when you discover the tag you accidentally left on the page promoting the coffee maker with actual smartphone sales statistics, the results are there: despite high product visibility, the product doesn’t sell well. The conversion rate is abnormally low.
So you’re looking for where the problem is … Is this the price to high? The description? The pictures? The conversion tunnel? In fact, it’s just an oversight … you simply did not change the tag designation from “smartphone” to “coffee maker” on the promo page.
Such cases can lead a product manager or site manager to take perfectly absurd actions: discontinuing the product, lowering its price, or even change the tunnel processing, could be costly in the long run.
Of course, the same process applies to online booking sites, configurators or advertising campaigns. E-commerce is not exclusive of these malfunctions.
To stay informed about the latest news Web Analytics and discover new case studies, please visit the blog Hub’Sales.