A few weeks ago, I wrote about cus­tomer ana­lyt­ics as the nat­ural next phase of life for many good dig­i­tal ana­lysts. (For more on cus­tomer ana­lyt­ics, see this post from yes­ter­day by my col­league, Adam Jenk­ins.) It was a sig­nif­i­cant depar­ture from my past writ­ings on this blog, most of which have been focused not just on digital-only ana­lyt­ics, but on prac­ti­cal items such as imple­men­ta­tion tips and prod­uct updates. I feel strongly about where we as an indus­try are headed, so at the risk of bor­ing those of you who are look­ing for Adobe Ana­lyt­ics tips and tricks, today I’d like to con­tinue shar­ing my thoughts and obser­va­tions around a par­tic­u­lar tenet of cus­tomer ana­lyt­ics which might be my per­sonal favorite: visual query.

What is visual query?

Imag­ine a con­tin­uum based on how eas­ily an ana­lyst can quickly and iter­a­tively probe their data for insights.

At the far left end are tra­di­tional query tools, like SQL. Ulti­mately, you can get what­ever you need, but after writ­ing a query (depend­ing on the size of the data set and other fac­tors) it can take hours or days get a result. If you have a follow-up ques­tion (which ana­lysts are great at think­ing up), you have to write another query and, once again, wait for it to return. This cycle con­tin­ues until the ana­lyst exhausts his or her patience and/or obtains the desired insight. But it’s painful to do analy­sis this way; it isn’t con­ducive to explo­ration at all. The ana­lyst thought process is more iter­a­tive than this.

In the mid­dle are a bunch of tools that seek to facil­i­tate bet­ter analy­sis in a few dif­fer­ent ways. For exam­ple, they might improve speed of iter­a­tion by apply­ing “struc­ture” around the queries that the ana­lyst is mak­ing. A tool like Ad Hoc Analy­sis (for­merly known as Dis­cover) in Adobe Ana­lyt­ics fits this mold. Dig­i­tal ana­lysts love Dis­cover because they can very quickly cre­ate new seg­ments, drill into their data, and cre­ate cus­tomized data tables and visu­al­iza­tions. The ana­lyst can move from thought to thought much faster than they can by writ­ing data­base queries man­u­ally, and the tool offers flex­i­bil­ity far beyond basic report­ing. But even here, the ana­lyst must still wait (albeit just a few sec­onds) for data to return. Fur­ther­more, a num­ber of restric­tions limit the way that the ana­lyst orders the behind-the-scenes queries that return their data to him or her.

As we approach the right end of our con­tin­uum, we are also approach­ing “speed of thought” for the ana­lyst. This is visual query—a con­cept where select­ing and explor­ing data natively changes the query on the fly, and results begin to return in real-time, even as the query con­tin­ues to retrieve data. Why do we call it visual queryBecause you are query­ing the data visu­ally rather than by writ­ing a state­ment, and because the data updates based on the query right there, visu­ally, across every­thing you’re viewing. The ana­lyst can do vir­tu­ally any­thing here; select­ing a row or a series of rows fil­ters the entire data set (what­ever data tables and visu­al­iza­tions are loaded up) by chang­ing the query. Build­ing a seg­ment or defin­ing a cohort is as easy, lit­er­ally, as click­ing a mouse. The ana­lyst is free to move from thought to thought lit­er­ally as quickly as they come to mind, to eas­ily move back up the “ques­tion tree” and pur­sue a dif­fer­ent branch at any time, and to really explore across mul­ti­ple dimen­sions, met­rics, and visu­al­iza­tions simultaneously.

Let’s say I’ve got four data tables open on my work­space. For sim­plic­ity, let’s pre­tend that they map nicely to reports that you’re famil­iar with: a Vis­its table (by day), a Prod­ucts table (show­ing Rev­enue), a Cam­paigns table (show­ing Rev­enue), and a Coun­tries table (show­ing AOV). In the Vis­its table, I sim­ply click and drag to select seven days. Instantly, the other three tables begin to update to reflect this selec­tion. Now I’m only look­ing at data for that one week. Now I select a par­tic­u­lar cam­paign the same way. Again, the tables update. That’s inter­est­ing; a prod­uct that I hadn’t expected to see has a lot of rev­enue tied to that cam­paign on those seven days. I select the prod­uct and de-select the cam­paign. Now I am look­ing at all the cam­paigns for users who pur­chased that prod­uct dur­ing those seven days. And on and on, across all of your dimen­sions and met­rics, cross-channel paths, any asso­ci­ated visu­al­iza­tions and more. Every­thing can be used to change your lens on any data that you’ve got avail­able to you, with mul­ti­ple pieces of data avail­able to you on-screen at once.

That’s visual query in a nut­shell, and it enables analy­sis like you’ve never done before.

Why is visual query so impor­tant to cus­tomer analysis?

We could actu­ally ignore the whole “omnichan­nel cus­tomer data” con­ver­sa­tion for the pur­poses of this post; visual query doesn’t nec­es­sar­ily require mul­ti­ple chan­nels of data in order to be tremen­dously valu­able to the ana­lyst, as I hope you can imag­ine. If you want to query your data faster and more intu­itively, visual query is for you—regardless of the data set you’re ana­lyz­ing. But in the world of cus­tomer analy­sis, visual query adds par­tic­u­lar value, and that’s worth dis­cussing briefly.

I met recently with a major U.S. retailer who has been think­ing about cus­tomer analy­sis. Since this is a well-known brand which has a tech-savvy audi­ence, it’s safe to assume that they have what most of us would con­sider “a lot” of web data. Still, accord­ing to them, web data rep­re­sents only about 15% of their total data store by size.  That means that 85% of their data store comes from other sys­tems and chan­nels: CRM, point-of-sale, call cen­ter, IVR, etc. But don’t focus on size; instead, think about how much more com­plex cus­tomer paths are than web vis­i­tor paths. That web path could be just a small frac­tion of the over­all cus­tomer path. As you’ve heard else­where, the cus­tomer jour­ney often nei­ther begins nor ends on your web site. As a result, the range of ques­tions that the ana­lyst will have as he/she moves through omnichan­nel cus­tomer data is far less lim­ited than in the dig­i­tal analy­sis world. We’re not just con­cerned about whether a user moved from Page A to Page B, or whether a vis­i­tor pur­chased Prod­uct X because she saw Cam­paign Y; poten­tially every inter­ac­tion that the cus­tomer has had with your brand, via any channel, is in scope. Thus, the abil­ity to iter­ate quickly, always revis­ing think­ing and ques­tion­ing, and to begin to get results imme­di­ately across mul­ti­ple views into the data becomes critical.

Think about it: You don’t want to be lim­ited by hav­ing to write queries and wait an hour or more for your next iter­a­tion, nor do you want to be roped in to even a flex­i­ble set of reports. Because you are now think­ing across the whole cus­tomer jour­ney, across mul­ti­ple chan­nels, the value of being able to explore and change your focus on the data (i.e., your query) with a sin­gle click in a data table, or on a col­umn in a visu­al­iza­tion, can become the dif­fer­ence between analy­sis in min­utes and analy­sis in hours or days.

Sim­ply put, visual query and cus­tomer data are a match made in heaven for the analyst.

What does visual query mean for how I should approach my world as a dig­i­tal analyst?

I still believe that cus­tomer analy­sis will become a big­ger and big­ger part of your world over the com­ing years, dig­i­tal ana­lyst. As such, insist on doing cus­tomer analy­sis using tools that allow you to query the data visu­ally. Adobe Ana­lyt­ics Pre­mium has this capa­bil­ity today, and as I men­tioned above, you don’t actu­ally need to wait to merge cus­tomer data into your web data in order to take advan­tage of the increased speed, effi­ciency, and flex­i­bil­ity that visual query pro­vides. You can enable more pow­er­ful analy­sis based on your web data today.

Ulti­mately, though, my goal is sim­ply to make you aware, in case you were not already, that your world as an ana­lyst will likely not revolve around basic, rigid reports in the near future as it may have done for many years now. In fact, the whole con­cept of a “report” changes when visual query is avail­able. Not to worry; I believe there will always be a time and a place for report­ing. Ana­lysts can have their cake and eat it, too. Think of visual query as in addi­tion to the stan­dard report­ing that your orga­ni­za­tion relies upon you to pro­vide. (Not to men­tion that the reports, seg­ments, audi­ences, etc. that you can cre­ate based on the results of your visual queries can help change the way your mar­ket­ing col­leagues use data.)

The future of ana­lyt­ics is visual. We see and hear this in a num­ber of places. We’ve all rec­og­nized for most of a decade that com­pelling visu­als help tell sto­ries, and telling sto­ries gets buy-in. I’m here to tell you that it isn’t just the out­puts of analysis—the Excel spread­sheets, the Pow­er­Point pre­sen­ta­tions, etc. that will be visual—the queries them­selves in the tools that you use will be visual, and the ana­lyst will be far bet­ter enabled because of it.