Adobe Summit 2017 — Major Announcements and What They Mean

Adobe Launch — The Future of DTM

On main­stage, one of the first announce­ments was about Launch, a complete rethink of Adobe’s Acti­va­tion core service (known today as Dynamic Tag Manage­ment, or DTM). Over the past four years, DTM has grown at a stratos­pheric rate, and Launch is an impor­tant mile­stone for several reasons.

First, Launch shows Adobe’s commitment to tag management.

Not only is Adobe putting dollars behind Launch, they’re also enlist­ing blis­ter­ingly smart people with real empathy for today’s marketers—and it shows in the soft­ware you saw at Summit. Adobe is putting some of its best teams and resources behind an impor­tant commit­ment to being a more open plat­form. Open­ness is crit­i­cal to marketers, and Adobe is seeing and react­ing to this (more on that later).

DTM is a crit­i­cal piece of the market­ing archi­tec­ture for every busi­ness who uses it, and the tag manage­ment space has seen a lot of compe­ti­tion. The dust has settled and there are a few clear winners, meaning it’s vital for busi­nesses to know that their choice will see contin­ued improve­ments, evolu­tions, and support. People have felt confi­dent in Adobe’s commit­ment to DTM, and the money and talent they’re putting behind Launch proves we should continue to be confi­dent in Launch.

Second, Launch is about openness, scale and customization.

Launch is API-first, allow­ing anyone to develop on top of it. This allows launch to hit three birds with one API stone:

1. For the first and only time in tag manage­ment, exten­sions and inte­gra­tions can be built by anyone. As prac­ti­tion­ers and consul­tants, we don’t have to rely on any single entity to build an inte­gra­tion for Doubleclick or Face­book, for Turn or Bluekai. Those compa­nies can build and main­tain their own, syncing with their own inter­nal devel­op­ment time­lines and new capabilities.The inte­gra­tions that have been built by tag manage­ment vendors in the past have always been limit­ing, and many clients avoid using them in lieu of writing their own custom versions or paying for one-off customiza­tions. But with Launch, anyone in the commu­nity can expand or write their own exten­sion if the vendor them­selves does not do a good job. If you have a partic­u­lar set of needs for track­ing embed­ded video content, for example, you have complete control.This allows for customiza­tion to be spread out over hundreds of people contribut­ing to the commu­nity, instead of relying on one TMS to try and keep with the hundreds of tech­nolo­gies they are designed to imple­ment.
2. Launch’s API-first approach also allows us to enhance the capa­bil­i­ties of Launch by writing func­tion­al­ity to help manage wide networks of sites at scale, rather than working at the web prop­erty level. We can deploy tech­nolo­gies across hundreds or thou­sands of sites, ensure that site or group-level customiza­tions are still intact where they need to be, and account for idio­syn­crasies and unique­ness across indi­vid­ual sites and appli­ca­tions. For the many global brands that work with us and have anywhere between 10 and 2,500 web sites, this func­tion­al­ity will create massive effi­cien­cies and improve­ments to gover­nance.
3. Third, devel­op­ers can now write sophis­ti­cated func­tion­al­ity to improve work­flow and manage­ment of Launch. For example, our own solu­tion, Sling­shot, allows marketers to use a single embed code across all of their DTM prop­er­ties, rather than having sepa­rate embeds for each web prop­erty, and staging and dev envi­ron­ments. For some of our clients, that means the differ­ence between manag­ing over 2,000 unique embeds and manag­ing 1 embed across thou­sands of sites.Similarly, we will be writing custom work­flows for approvals, alerts, status, and who needs to do what. After hundreds of engage­ments at Search Discovery, we’ve found these to be the best prac­tices for imple­men­ta­tion and main­te­nance. Launch will work exactly the way we and our clients need it to, allow­ing us to make the most of our engage­ments and focus on strate­gic work that moves KPIs.

Third, geekiness.

Now, this is the part that some people are excited about, and that others fear. While the devel­op­ers out there are doing back­flips over Launch, many busi­nesses are wonder­ing, “Does what I just saw mean that DTM just got too tech­ni­cal for us to handle?”

I have some pretty strong thoughts about this topic, which I’ll illus­trate through a story.

When Search Discovery CEO, Lee Blanken­ship, and I were first start­ing the Satel­lite busi­ness, before it was acquired by Adobe and became DTM, we had a meeting with one of the 25 biggest compa­nies in the United States. Their primary crite­ria for choos­ing a tag manage­ment system was that “even a completely non-tech­ni­cal person can use it.”

I’ll say now what I said then: “For me, no matter how easy a Boeing 747 ever becomes to fly, at no point will I ever be okay with a non-pilot flying it. And I think your busi­ness should have the same atti­tude about who’s pilot­ing your data collec­tion across 100% of your digital exis­tence.”

Now, please note two things: 1) we fortu­nately won that busi­ness and had a massively success­ful deploy­ment, and 2) I am not in any way saying that these tools shouldn’t strive to smarter, safer and easier to use—that is always a good thing.

I’m simply and adamantly point­ing out that data is the lifeblood of busi­ness. In most busi­nesses, dozens of plat­forms, automa­tion systems, analy­ses and insights rely completely on good data, and at no point should anyone consider handing the controls over to someone without the requi­site skillset. You are not staffing that pilot’s chair for the sunny, calm days when every­thing goes exactly as planned. You’re staffing that chair for the storms, and every busi­ness has storms. Most have them almost daily.

Make sure you have the right people in these crit­i­cal roles, whether that is in-house, or through a proven partner. Please, do not make a mistake on who the pilots are.

 

But the question still remains: Is Launch now too technical?

The answer is “no.”

In reality, the parts of DTM that needed to be more tech­ni­cal to allow amazing people to do amazing things have opened up. The parts of DTM that were user-friendly have also gotten better at the same time. In a sense, DTM, as it evolves into Launch, is getting both more and less tech­ni­cal at the same time, and this is a great thing.

This allows groups like us to do things with greater effi­ciency and scale than ever before, while allow­ing less-expe­ri­enced teams to handle simpler use cases with greater ease of use, orga­ni­za­tion, find­abil­ity, safer and more nuanced approval work­flows, and more.

Getting into the nitty gritty of Launch is some­thing we don’t have the space for here, but believe me, the future is bright and we are looking forward to helping our clients make the most of the oppor­tu­nity. Stay tuned to our blog for much more infor­ma­tion as Launch gets closer to…umm….launching.

 

Let’s talk about what the Data Platform is/could be.

Adobe’s Expe­ri­ence Cloud (Market­ing Cloud) archi­tec­ture is made up of many acqui­si­tions, in addi­tion to home­grown tech­nol­ogy. Each of these pieces of product and func­tion­al­ity were at some point free­stand­ing busi­nesses, each with their own data collec­tion, user admin­is­tra­tion, data process­ing, algo­rithms, report­ing inter­faces, segmen­ta­tion tools, etc. And perhaps most impor­tantly to the way marketers work today, they each had differ­ent data storage systems and schemes.

The impact of this is easily illus­trated by this example. Compare the number of visits, clicks, page views, conver­sions, members of segments, etc. measured by each solu­tion (Analyt­ics, Target, Media Opti­mizer, Campaign…), and you’ll find that conver­sions, as well as just about every other metric, are going to be differ­ent.

This is the point at which people start to freak out a little bit. Let’s not go down that road. The team here at Search Discovery will certainly continue to create some diag­nos­tic content and offer services address­ing when it is and isn’t okay for numbers to be differ­ent, but I’d like to focus on some­thing much more strate­gic here.

Looking beyond the differ­ences in the numbers, the most impor­tant topic to modern marketers is the ability to manage their market­ing and behav­ioral data holis­ti­cally. What wins in digital today is not prod­ucts, but archi­tec­tures. It’s not about which email plat­form is best in a silo; it’s about which email plat­form works best in your overall tech­ni­cal and strate­gic ecosys­tem: being able to tap into any cross-channel audi­ence or set of trig­gers to inform the topic, timing, and desti­na­tion of those emails, for example (right message, right time, right person, right place).

The process of buying and imple­ment­ing tech­nolo­gies in silos is already dead at the most success­ful and advanced busi­nesses. The vital signs say the same for the broader adop­tion curve—“must play well with others” is the funda­men­tal require­ment of the future.

The data plat­form that Adobe is working toward provides a multi­tude of bene­fits. Here are a few that I think we can all start dream­ing about today:

• Collec­tion: A single data plat­form paves the path for a single collec­tion pipeline for all behav­ioral data—one archi­tec­ture to main­tain, opti­mize for perfor­mance, and handle every customer touch point. One customer iden­ti­fi­ca­tion system (which we’re already seeing with the visitor ID service) tying all behav­ior together at collect time, meaning many data marriage exer­cises later in the process are easier or unnec­es­sary.
• Storage: Want answers? Get all of them from one place, regard­less of the product mix you are using. This logi­cally leads to central­ized report­ing and analy­sis for all forms of data Adobe manages and imports, and one place for you to query on your own, as those tools hypo­thet­i­cally come avail­able. Also, if you want to export into your own data warehouse/lake/cloud/galaxy/ball of yarn (or what­ever you’re calling it in your busi­ness), this leads to the possi­bil­ity of getting the whole Adobe batch, all clean as a whistle across all web and app analyt­ics, search, display, social, person­al­iza­tion, email, video, etc. at once. That way, there’s no struc­tural work piece­meal on your side.
• Inte­gra­tion: No need for solu­tion-to-solu­tion inte­gra­tion of events, trig­gers, rules, models, audi­ences or stored data. It all lives in the same home and gets along. Exter­nal inte­gra­tion is central­ized, so any 1st or 3rd party data sources or appli­ca­tions inter­face in one place.
• I/O: A single schema for the import and export of all types of data, which could logi­cally also lead to a true stan­dard for these types of data (yes!).
• Appli­ca­tions: Complex cross-channel + cross-device + multi(and stitched)-session analy­ses, predic­tive algo­rithms, rule sets and audi­ence manage­ment, attri­bu­tion, and the list goes on and on. Is there some­thing you’d like your data universe to reveal or trigger? That is going to be a whole lot easier.

This is just a sample of the outcomes. Are you start­ing to appre­ci­ate how big of a deal this is? Adobe has fully acknowl­edged that archi­tec­tures trump silos, and the data plat­form could­n’t be a clearer signal of the new era they are walking through the door. Most of the above is just me think­ing through (and hoping) what it will turn into, so it could take on much more as you use your own imag­i­na­tion.

And that brings us to the second reason this is inter­est­ing:

 

Let’s extrap­o­late this idea: if many acqui­si­tions brought over many data pipelines and storage approaches, it stands to reason that they also brought numer­ous client-side imple­men­ta­tions, edge services, scoring algo­rithms, report­ing inter­faces, chart­ing APIs, report automa­tion systems, alert­ing systems, activ­ity logging, user manage­ment, asset stores and managers, deci­sion­ing systems, rule builders, etc.

Start­ing to get the picture? It doesn’t take too huge of a mental leap to under­stand that this whole ecosys­tem is headed in a revo­lu­tion­ary direc­tion that will have more bene­fits than we can count. Abstracted, gener­al­ized compo­nents (includ­ing and beyond the Core Services offered today) allow a single archi­tec­ture to do many, many things, and do those things with more effi­ciency than mono­lithic, free­stand­ing appli­ca­tions can.

 

The last piece of main­stage content wasn’t about product, but putting their brand­ing where their hopes are, and that is posi­tion­ing the Market­ing Cloud as the Expe­ri­ence Cloud. There was also talk about the Analyt­ics Cloud and the Adver­tis­ing Cloud.

We had many visi­tors to our booth ask why there were clouds of clouds on clouds, and the best reply I could come up with was that “Stacks on Stacks on Stacks” was already taken, and that’s what Adobe is trying to do for you: make you a whole bunch of green­backs.

What I believe is impor­tant here is not how this repre­sents the prod­ucts differ­ently, but how this repre­sents Adobe’s intent and the way they would like for us all to be think­ing: this isn’t about Market­ing, it’s about Expe­ri­ences. All of it. We are creat­ing, mone­tiz­ing, opti­miz­ing, person­al­iz­ing, and enrich­ing expe­ri­ences.

I believe and hope it’s a much more authen­tic repre­sen­ta­tion of what brands will be striv­ing for in this partic­u­lar era.

I, for one, want to be a consumer that brands care about enough to archi­tect great expe­ri­ences. They will get more of my money, for sure. But more on that in the last section of this overview, “How to commit to getting better at digital expe­ri­ences.”

Next: What we heard from over 400 conver­sa­tions with real analysts and marketers at Adobe Summit