The Future of Digital Marketing

Age of performance marketing

Marketing gamechanger

When I started learning marketing at university, I didn’t know how far it had been digitized. I thought all you needed was an outstanding campaign, and the results would have reflected how good that campaign idea was.

Google and Facebook ads have changed that. The code tracks everything from the banner views up to purchases. You can now tell how many clients you’ve acquired from a specific ad. This enables the ROI machine. When you are confident that a particular channel returns 5x of your expenses, you’ll increase the budget for that channel.

Performance marketing is a prime example of science manifestation in marketing.

Growth engines & metric calculating

Performance marketing is so effective it has been identified as one of the core growth engines of business, as told by Eric Ries in The Lean Startup. The company will grow as long as lifetime value (LTV) exceeds customer acquisition cost (CAC). It’s incredibly convenient for businesses with recurring purchases like eCommerce and SaaS.

I love performance marketing because it’s as easy as keeping the scores. In our space, we have a term for planning performance metrics – unit economics. The idea is to calculate every performance metric to make financial sense. Before the business launch, unit economics are used to determine if a business model will be profitable. After the launch, unit economics optimize performance metrics like CAC and conversion rates (CR).

Tracking offline activities

Even the offline marketing channels have integrated with performance tracking:

• Out-of-home (OOH) billboards include QR codes to count the website visits

• Retail stores use proximity marketing tech like geofencing, Bluetooth beacons, and NFC tags

With that many channels, you’d need a funnel nurturing approach that applies to all channels simultaneously.

What are some ways we can digitize the offline marketing channels?

Onwards to omnichannel marketing

Total data control

Omnichannel analytics is when you track every touchpoint between your brand and the target audience. “Omni” is an interesting choice of words. I associate it with “omnipotence,” a power so unlimited that anything is possible. No wonder it’s usually applied to gods.

In marketing, omnichannel analytics emits the same vibes. Imagine opening your GDPR data files on Netflix and seeing every tiny occasion you’ve interacted with Netflix:

Besides total privacy breach creepiness, you’d also be amazed by how magical this technology is, omnipotent even :)

Current omnichannel efforts are focused on building accurate data dashboards. The goal of these dashboards is to merge the funnel with the customer journey. At the top of the funnel, you’d have your brand awareness channels transitioning into acquisition and retention channels.

Do you think companies will share their data with each other to build complex user persona?

Omnichannel marketing is huge email marketing

Another way I like to describe omnichannel marketing is “email marketing gone big.” Because in email marketing, you have a sequence of automated emails based on the current funnel stage. In the trial stages, you’d be sending emails to convert into paid users. And then for paid users, you’d send emails to convert into retained users.

The same pattern occurs for omnichannel marketing. The difference is, the sequence is now made of messages across all channels. If you hit your acquisition goal on Google Display Ads, you can directly target on Facebook Ads to convert the acquired users into paid ones. Here’s a more detailed look at omnichannel marketing.

Omnichannel marketing limitations

Sadly, omnichannel marketing is limited in tracking brand awareness attribution. A typical way to measure brand awareness is by monitoring the monthly search queries of a brand keyword. Basically, how many times your business name has come up in Google searches. But you can’t attribute the growth of this metric to a specific channel. Moreover, search queries only represent a share of all people who are familiar with your brand.

A similar challenge occurs when performance marketing clashes with PR. You can count the views per article, but you can’t know precisely how many new people have learned about your brand’s existence. One way to deal with this is to stop obsessing about these numbers. Or embrace the probabilistic models.

Should we leave awareness and PR campaigns alone from performance marketing?

Online privacy trends

Privacy vs marketing

It appears that tracking people for selling products to them is like being a stalker! It makes you think about “targetting” in a different light.

Indeed, CRMs like HubSpot enable us to track the web actions of a specific person. It’s just necessary to build that omnichannel marketing process. But it also allows these CRMs to become single points of privacy failure.

If only we could aggregate data without revealing identities, huh? Well then how would you build the account-based marketing flow? Or for that matter, any performance marketing in B2B?! Privacy regulations seem to disrupt performance and omnichannel marketing.

Google updates

Google Chrome is set to stop supporting third-party cookies soon. I am writing soon because this is their second time postponing this action. Blocking third-party cookies is a big deal because that’s what powers your Google Analytics and other adtech on the websites.

It turns out third-party cookies are actually one of many technologies to track users. The privacy updates are meant to enforce consent from a user, not remove tracking altogether.

Apple updates

Apple outran Google by at least 4 years now. Safari has started blocking 3rd party cookies since 2020. And iOS 14 has blocked acquisition attribution by blocking IDFA (Identifier for Advertisers). IDFA is used to identify devices.

Without IDFA, marketers need to rely on fingerprinting – gathering publicly available data on a mobile device to construct a single device data profile. But fingerprinting is just not as accurate as using IDFA.

Brave – crypto outsider

One outsider in particular has decided to reinvent the browser’s use for marketing – Brave. The browser is one of the old-school crypto projects emerging in 2016. The idea is to cut ad networks from monetizing websites. To do this, Brave has a built-in ad blocker.

Basically, advertisers purchase BAT tokens to spend them on ads inside Brave. The ad viewers then receive a part of these spent tokens. The users can then decide to reward frequently visited websites with token donations. I suppose in an act to say sorry for blocking your ads :)

How can we still market to people without breaking their privacy?

Probability models instead of real data

Emerging trend: probabilistic models

Before technology or culture goes mainstream, it has to trend in a niche community. In our case, my bets are on mobile app marketing. In this space, apps rely on Mobile Measurement Partners (MMP). MMPs are attribution engines – they match acquired users with their source channel to improve campaigns’ ROI.

iOS 14 has made MMPs’ jobs more critical for performance marketers. Since it has become harder to identify a specific device, MMPs have started using probabilistic models. Probabilistic models use temporary data that becomes obsolete within a few hours:

• device info

• time of install

• time of click

For probabilistic algorithms, this data is enough to estimate the source of an install for a few hours after a click. The estimations are accurate enough to optimize campaigns based on the data.

Improving conversions based on model data

Probabilistic modeling also applies to conversion rates! MMPs analyze users who have consented to be shown personalized ads. Then, they extrapolate their post-install behavior to model aggregate user data. Basically, it’s suggesting data for opted-out users – based on data for opted-in users.

Somehow, this can also work for targeting opted-out users.

Probabilistic modeling going worldwide

If probability modeling resolves privacy issues, it might become mainstream. But we’d still need a fair share of people who consent to share their data. Because that’s going to be the input for all our models. In this case, it’d be wise to incentivize opt-ins. Current incentives are only verbal and are based on brand trust:

• “You are in control of your data”

• “We’ll show you relevant ads”

It would be so odd and yet interesting to see Google Analytics showing confidence levels and approximate models, not real data. And HubSpot would show a probability percentage of a contact’s acquisition source.

Only time will tell if this trend becomes mainstream.

Do you think it’s possible that we’ll make a wide transition to model-based analytics?

Subscribe for more conversations!

There’s more where that’s coming from!
Subscribe to the newsletter to receive a new article every 2 weeks: