Marketing AI and Machine Learning Come Into Better Focus

Artificial intelligence and machine learning are two phrases that have become regular currency in the marketing world over the past several years. It isn’t hard to figure out why, as both AI and machine learning have the potential to help marketers make sense of the ever-increasing volume (and complexity) of raw data that’s become available in increasing amounts, thanks to the digitization of “everything.”

Some people use the two terms interchangeably, but that isn’t exactly right. According to Thorin McGee, director of content at Fast Capital 360, the distinction is subtle yet significant:

  • AI is when you develop an algorithm that allows a computer to “think” for you towards achieving a goal.
  • Machine learning is when you let a computer create an algorithm to find ways to meet the goals you give it, based on large pools of data.

Put the two together, and you have the ability to gain some really deep insights into what your data is actually telling you, thereby improving decision-making success.

On the data front, this great potential is tempered by some significant challenges. Christopher Penn, chief innovation officer of marketing data and analytics consulting firm Trust Insights, characterizes them as the “5 V’s” of data:

  1. Volume — There’s so darned much of it.
  2. Variety — More kinds of data are being churned out.
  3. Velocity — Data is coming at us faster than ever.
  4. Veracity — If data isn’t verified, it can do more harm than good.
  5. Value — In raw form, data isn’t particularly useful. Like oil, data needs to be refined to be of value.

If getting your arms around data seems like trying to hug a stream of water, you aren’t alone in thinking that. Many companies are pretty adept at using data to identify what happened — and maybe even to diagnose problems and why they happened.  But it’s less easy to predict what will happen based on data … and even harder to use data to determine with confidence what should happen.

The biggest challenge — but also the one with the biggest potential payoff — is tapping machine learning to process and use data in forging future business as you wish it to be.

To date, very few companies have come all that close to becoming AI-powered enterprises. But it’s where we’re headed in the coming decade.  It represents one of the biggest opportunities for differentiating one company from another.  But it will require a disciplined and concerted effort:  talent acquisition (developers and data scientists), tapping outside vendors, along with taking available open-source code and building upon that to implement the appropriate marketing technologies.

Oh, and committing to a multi-year initiative and budget even after all of those other pieces are in place.

Surveying the current landscape, are there particular entities that you see as on the leading edge in applying AI and machine learning to their marketing endeavors? Please share your observations with other readers.

Any way you slice it, Google continues to dominate the search ecosystem.

Just how big is Google in the world of search? I’ve seen percentages that are all over the map, but one thing is undeniable:  Google remains the overwhelming leader in search.

And it isn’t even close. Runners-up in the search engine derby include Bing/Yahoo and DuckDuckGo, but they’re so small so as to be mere asterisks at the bottom of the page.

… Which might be surprising to some. After all, as late as 2015 comScore was reporting that Google’s market share of desktop search was running around 64%, whereas the Bing family of search products (including Yahoo and AOL) was tracking in the low 30s.

But it’s all in how you make the calculations. At the very same time, Statista was reporting that Google’s worldwide share of desktop search was approximately 89%.

Moreover, Statista’s trend line for Google between 2010 and the end of 2018 is remarkably consistent, with Google’s share of desktop search charting in a narrow range between 86% and 90%:

But I think it’s the data from marketing intelligence and analytics firm Jumpshot that gets us closest to what’s actually happening in the world of search. Jumpshot licenses anonymized ClickStream data from hundreds of millions of users.  It finds that ~63% of all online searches are through Google’s “core” function.

But then one needs to factor in additional Google-related search activity that occurs on Google Maps, Google image search and YouTube, which is owned by Google. When those figures are added to the mix, Google’s market share of search is indeed in excess of 90%, with all other players way, way behind.

This graph shows the makeup of Google’s dominant position as compared to its search competitors:

Source: Jumpshot (based on ClickStream data), 2018.

These dynamics explain why Google remains so entrenched – and why advertisers continue to devote so much of their search engineering advertising dollars to Google properties.

A “constant” in Google’s market strategy over the years has been to make it easy and effortless for users to perform a Google search wherever they are.  In years past, that meant making Google the Home Page on as many Internet browsers a possible. In more recent times it’s taken the form of building activity on Google-centric browsers (Google Chrome), mobile market share (Android), acquiring the dominant video platform (YouTube), and making a major push into voice search with Google Home.

Essentially, wherever someone is … Google is there as well. It’s very much like a commodity or a utility.  (Indeed, its very name has become synonymous with the verb “to search.”)

In case anyone is counting, Google processes an eye-popping 3.5 billion searches per day.  Is it any wonder that any competitor – even a platform like Bing with resources to spend – would have a near-insurmountable challenge getting millions of people to just try a different search option (much less start using it regularly).

Could the situation change?  I suppose nothing is immutable.  The market share figures don’t yet factor in iPhone data at scale.  Some other search product might emerge that is dramatically better-performing than Google.

But none of those factors are likely to change the overall search ecosystem. The fact is, Google dominates search … it has dominated search for years … and it’s on track to continue doing so in the future.

I close with a question to readers. If any of you prefer using a different search product besides Google, please share your reasons why in the comment section below.

Chief Marketing Officer: The most thankless job in the corporate world?

Few people I know would claim that being the Chief Marketing Officer of a company is a job without risks. Indeed, numerous articles in the business press point to an average length of tenure in a CMO position that is often measured in months rather than in years – indeed, the shortest length of time among all C-level jobs.

And now, a recently completed survey of CMOs  underscores just how wide-ranging the reasons are for those employment characteristics. Branding consulting firm Brand Keys tested a number of issues to see which are the ones that keep CMOs “awake at night.”

Three-quarters or more of the respondents to the Brand Keys survey reported that every factor presented was significant enough to cause them to lose sleep.  Leading the list with near-universal high-alert concern is ROI factors. Other factors of concern to nearly every respondent in the survey are big tech and data security issues.

Listed below is how each of the factors tested by Brand Keys turned out with CMOs in terms of “losing sleep” over them.

90%+ lose sleep worrying about:

  • ROI/ROMI factors
  • Big data, big tech and big security issues
  • Establishing trust with customers
  • Innovation, AI, technology and marketing automation developments
  • Consumer expectations regarding privacy and transparency

80%-90% lose sleep worrying about:

  • Managing social networking
  • Creating relevant advertising content
  • Successfully deploying predictive consumer behavior analytics/technologies
  • Dealing with consumer advocacy and social activism
  • Developing long-term strategies that align with corporate growth goals
  • Having the ability to engage with audiences – not just find them

At the “bottom” of the pile … 75%-80% lose sleep worrying about:

  • “Democratization” of the digital world and protecting brand equity within it
  • “Political tribalism” and its effect on brand reputation
  • Being relevant when tweeted about
  • Keeping consumers engaged with the brand
  • Creating better cross-platform synergies for marketing campaigns
  • Creating an “unlearning curve” to move away from legacy marketing metrics
  • Creating marketing synergies among different generational/age cohorts
  • Being replaced by the chief revenue officer

This last worry factor – losing their job – seems almost preordained given the tenuous circumstances more than a few CMOs deal with in their positions.

… and likely made more so because CMO’s are quick to be blamed when things don’t go well, even if they aren’t in the strongest position to effect the changes that may be needed. “Responsibility without authority” is the stark reality for too many of them.

What are your thoughts about the dynamics faced by CMOs in their companies?  Whether you speak from personal experience or not, I’m sure other readers would be interested in hearing your views.

 

No End in Sight to the Challenge of Email Deliverability

When it comes to e-mail communications in the B-to-B world, yet another study is underscoring just how challenging it is to reach corporate inboxes.

A new report by cyber-security firm FireEye, Inc. reveals that fewer than one-third of e-mails sent are actually making it into corporate inboxes. The FireEye analysis was based on tracking more than a half-billion e-mails sent between January and June of 2018.

The majority of those e-mails were deemed to be spam or malicious in their intent. Nearly 60% were blocked by threat intelligence and around 10% more were halted by attack prevention tactics such as URL inspection and attachment detonation.

E-mails were deemed suspicious because they triggered one or more of the following “red-light” cautions:

  • Malware-less impersonations
  • Malware viruses
  • Phishing attacks
  • Ransomware
  • Spyware
  • Trojan horses
  • Worms

Interestingly however, it turns out that only a small fraction of the e-mails actually had malicious intent, meaning that the super-strict filters being employed by companies are capturing a huge number of perfectly legitimate e-mail messages in their dragnet and rejecting them out of hand.

On the other hand, the FireEye analysis also determined that impersonation attacks have undergone a shift from domain name spoofing to “friendly” domain name scams – ones in which an e-mail address is manipulated to impersonate a trusted source.

As the study cautions:

“This shift in tactics may be driven by how easily cyber criminals can ‘spoof’ the display name and username potion of an e-mail header. Instead of having to go through the process of buying and registering a domain similar to – or one that sounds like – the recipient’s domain, they can simply change the display/user name.”

The FireEye analysis is a reminder that because of its sheer pervasiveness, e-mail communications are also the most popular conduit for potentially significant cyberattacks. No wonder companies have their guard up.

The problem is, clearly a whole lot of wheat is being thrown out with the chaff.  And that makes e-communications hardly the slam-dunk communications tactic that many people assume it to be.

Is third-party marketing data on life support?

As a marketing professional for the better part of four decades, I can’t imagine any of us doing our jobs without soaking up as much data as possible to help with our decision-making.

And data accessibility is miles ahead of where it was when I first entered the marketing field.  Back in the day, “finding data” meant hitting the reference libraries to access government or other reporting – especially if you were lucky enough to be located within a reasonable distance of one.

There was the phone for real-time information-gathering … and also the FAX machine for quick receipt of “facts in brief” — not to mention the “wait-and-wish-for” mail and package delivery services.

If it was insight you needed from customers or prospects about a new industry or business venture, primary research was always an option — if you had the money and the time to allocate to the effort.

As for “first-party” data, that was available as well – but how often were we at the mercy of the bureaucratic machinations of in-house IT departments to get even basic data requests processed in a timely way?

All of which is to say that marketers have always used data – but the quantity wasn’t as great, while the timeframe of data acquisition was at a snail’s pace compared to today’s reality.

But now, after having become quite spoiled at the availability of all sorts of information, might it be that we’re regressing a little?

In particular, third-party information purchased in bulk, often from data aggregators, seems to be where the backsliding is occurring.

Consider ad targeting and building audiences: We have access to valuable first-party data thanks to website analytics and studying the results of our own e-mail campaigns.

There’s no question that the first- and second-party data which marketers are able to access are highly valuable in that the information creates efficiencies in campaigns and drives higher conversion rates. But theoretically, the ability to layer on accurate third-party data would take things even further.

There’s also been third-party behavioral data from three big behemoths — Google, Facebook and Amazon – that can be used for MarComm targeting purposes. But of those three platforms, just one of them allows third-party data to be made publicly available to end-users.

This poses challenges for the suppliers that aggregate and sell third-party data, as the quantity and quality of their information isn’t on the upswing at all.

Fundamentally, finding a good source for third-party data entails understanding what sources each data aggregator is using and the methodology it employs to collect the data.  Factors of scale, quality, reputation and price also come into play.

But despite best efforts, when testing third-party data for MarComm campaigns and lead-generation efforts the results are often pretty ugly — the data loaded with inaccuracies and basically terrible for efficiency metrics.

It doesn’t help that with the rise of Amazon as yet another “walled garden” of data, the “open web” represents a ever-smaller portion of the total ad spend — and hence also a decreasing amount of the third-party data that’s available to end-users.

With the veracity of third-party data becoming more suspect, it’s had an interesting effect on data management platforms, which are now focusing more on the actual messages themselves and not the “personas” of the people receiving the messages or how they were identified and targeted.

Is it possible for third-party data to provide good information to AI systems — intelligence that can verify and augment the value of the first-party data? If leading ad platforms can use such third-party data to enhance the accuracy and value of what they sell to advertisers, there still may be valuable material to work with.  As it stands, though, I’m not sure that’s the case.

What are your experiences?  Please share your perspectives with other readers here.

Programmatic ad buying in the B-to-B sector: The adoption rate grinds to a halt.

Each year, Dun & Bradstreet publishes its Data-Driven Marketing & Advertising Outlook report.  The report’s findings are based on a survey of marketers in the business-to-business sector.  Among the questions asked of marketers is about the advertising tactics they utilize in support of their sales and business objectives.

A look at D&B’s annual outlook reports over the past several years, an interesting trend has emerged: The adoption rate of B-to-B companies being involved in programmatic ad buying has plateaued at somewhat below 65% of firms.

In fact, you have to go back to 2015 in D&B’s reports to find the proportion of companies involved in programmatic advertising running significantly below where it is now.

That being said, those firms that are involved in programmatic ad buying are planning on allocating additional funds to the effort. The most recent survey finds that ~60% of the respondents involved in programmatic advertising plan to increase their spending in 2019.  That includes ~20% who plan to allocate a significant dollar increase of 25% or greater.

Another interesting finding from the 2018 survey is that there appears to be slightly less interest in display and video programmatic ad placements – although display remains the most commonly run ad type.

Where heightened interest lies includes one category that should come as no surprise – mobile advertising – as well as several that might be more unexpected. Social media advertising seems like it wouldn’t be a very significant part of most B-to-B ad buyers’ bag of tricks, but two-thirds of respondents reported that programmatic advertising in that sector will be increasing.

Another interesting development is that ~17% of the respondents reported that they’re stepping up their programmatic buying for TV advertising – which may be an interesting portent of the future.

Lastly, the survey revealed little change in the types of challenges respondents face about programmatic ad buying – namely, how to target the right audiences more effectively, how to measure results, and the need for better technical and operational knowledge for those charged with overseeing programmatic ad efforts inside their companies.

More information and findings from the 2018 D&B report can be viewed here.