As we celebrated the recent launch of our Martech Intelligence (MIR) report on the price of business marketing and predictive analytics, I thought about the effectiveness data. I find it fascinating that our industry is constantly raising the bar when it comes to measurable and innovative barriers, such as new privacy rules and increasingly weaker cookies. That’s one of the things that gives me the most energy to work in this space – it never ends.
And if that doesn’t stop, Tuesday is the first newsletter with our new research focus. Each week I collect important and informative research that we have recently discovered and we will also conduct our own research and ask our audience for their opinion on topics that challenge the industry, from strategic thinking to current best practices. What you say and share the results so you can learn from each other – we hope you will participate and draw valuable conclusions in your day-to-day work.
This week we asked you to tell us how you replaced the Martech battery in 2021. We are very curious how the events in 2020 have affected your workflow.
Read on to learn more about how researchers’ devices affect your ability to click on the first results and see trends in the cell analysis space.
The real-world used for synthetic data
There are privacy considerations when it comes to personally identifiable information (PII). One solution was to make data records anonymous before they were analyzed. Another solution is to use an algorithm to “synthesize” data online, creating synthetic data that cannot be attributed to real individuals. Let’s be clear: Synthetic data is not as accurate and reliable as real data, but it has important uses for marketers.
One use case is precisely to bypass regulations such as GDPR, such as PII, the real data that identifies real people. Synthetic data can be obtained, stored, and managed in ways that may not be possible under GDPR regulations. Other uses include the need to obtain data to extend the actual data sets. For example, it can be used to extend datasets with predictive analytics or training data needed to train AI models. It can also be easier to generate synthetic datasets without sacrificing AI training.
There are cases where synthetic data can be just as useful as real data, such as training models that use autonomous mechanisms or extend data sets for in-depth image processing. These are not cases where the goal is to understand and predict consumer behavior. “Important data will never be as accurate as real data,” said Maciej Pondel, researcher and machine learning specialist at Unity Group in Warsaw. Even when synthetic data are generated from real models, they always lose the essential “reality factor”, making them useful in a limited number of cases.
Applying synthetic data to digital marketing is an evolution, not a revolution. Questions are limited and focused on emergencies. This will be another function in the toolbox.
Connections are proving to be extremely important in the pandemic
Even before the coronavirus pandemic disrupted Americans’ lives and businesses, the phone and especially the smartphone became an integral part of consumer purchases. More than 170 billion mobile calls for US companies are expected by 2020, according to BIA Advisory Services.
During the worst public health crisis in more than a century, American consumers have used the phone more than ever. In the first half of 2020, Google My Business grew 61% in the number of calls from consumers, from consultancy to billing contracts. In the twelve months ending in June 2020, calls were up 35% and are still 27% above the pre-pandemic level.
New perspectives on the eternal question
Specialized search ads are constantly optimizing their campaigns to get the best results at the lowest cost, and new research published in the Journal of Advertising Research (JAR) explains the importance of a critical factor: the device on which the ad word appears. The research found that the devices that people use when searching have a big impact on the ability to click or scroll based on the first result, at least as far as ticking is concerned.
Tablet and smartphone users previously clicked on the desktop in the first generic search result. When comparing tablets and smartphones, it is interesting to note that people using tablets are more likely to click on the main result for common searches, confusing the interpretation that smaller devices are more dependent on better results. In the search for brand, the attractiveness of the best result remained stable for tablet users, but no significant difference was found between computer and smartphone users.
The results add weight to what has become a best practice proposal for search engine advertisers: use bid modifiers and even split devices across different campaigns, allowing you to customize channels based on the characteristics of your search engines. For example, you can compete aggressively for first place in campaigns that work on tablets and in campaigns that are not used by brands and run on smartphones.
The study’s authors, Chongyu Lu from Pace University and Rex Yuxing Du from the University of Houston analyzed more than 20 million ad impressions from 13 different advertisers across multiple industries.
On my way
Confluent, the event streaming platform, announced yesterday that Stephanie Buscemi will join the company as CMO. Buscemi was the youngest CMO at Salesforce, a position she held for two and a half years and led a team of 1,400 marketers. He spent a total of more than six years in marketing roles at Salesforce.
In January, when Buscemi announced her departure, Salesforce promoted EVP and GM, Platform & AppExchange, Sarah Franklin, to the CMO role.