Interview with StoneTemple Consulting

Below is the tran­script of an inter­view I recently did with Eric Enge of StoneTem­ple Con­sult­ing.

Eric Enge: Let’s start by talk­ing about a basic overview of Omni­ture, and then move into an overview of Omni­ture SearchCenter.

Bill Mungo­van: Omni­ture focuses on online busi­ness opti­miza­tion and is the largest soft­ware com­pany focused on the CMO. Our Online Mar­ket­ing Suite includes our web ana­lyt­ics tool, Site­Cat­a­lyst. It also includes 9 other prod­ucts, such as Gen­e­sis, an inte­gra­tion tool that pulls in data from other sources and Test&Target, a land­ing page opti­miza­tion and a mul­ti­vari­ate test­ing tool, and of course SearchCenter.

Eric Enge: Is Test&Target based off the acqui­si­tion of Offermatica?

Bill Mungo­van: Yes, Test&Target is basi­cally based off of the Offer­mat­ica tech­nol­ogy. It is a dynamic Land­ing Page Opti­miza­tion with Mul­ti­vari­ate Test­ing on its land­ing pages.

Search­Cen­ter was basi­cally built in the con­text of that mar­ket­ing suite. Search­Cen­ter is what we call a search man­age­ment tool, in that it accesses each of the major search engines from a sin­gle loca­tion and pro­vides auto­mated bid man­age­ment and port­fo­lio opti­miza­tion. You can access all sorts of dif­fer­ent report­ing func­tion­al­i­ties through the Site­Cat­a­lyst integration.

We think about Search­Cen­ter as a tool for search mar­keters, but given the fact that we have Gen­e­sis, we can pull data in from other sources, like an email provider, an ad-server, a CRM sys­tem like Sales​Force​.com or a client’s cus­tom, inter­nal database.

Eric Enge: Right. Pulling in data from other sources is one of the big chal­lenges with bid or cam­paign man­age­ment. Peo­ple are so used to treat­ing every­thing like they are direct response mar­keters. But a com­pany that has phys­i­cal loca­tions, and a web site, is likely going to have inter­ac­tions with peo­ple going to the web site and buy­ing offline, and vice versa. So, being able to pull in data from other sources allows you to credit those cam­paigns in a mean­ing­ful way so that you can more effec­tively man­age your bid­ding strategy.

Bill Mungo­van: Yes. That’s the heart and soul of the way we think about search, which is obvi­ously the hot topic. We use Site­Cat­a­lyst to col­lect all of that data. So, in your exam­ple, you could pull in point-of-sale data or data from a call cen­ter. There is really no short­age of exam­ples there. Then we can gen­er­ate bid rules and bid strate­gies based on that data.

That’s how Omni­ture thinks about the world, given the fact that we have Site­Cat­a­lyst as an under­ly­ing plat­form. We can pull data in from all these dif­fer­ent sources, and then use that data not just for attri­bu­tion, but also to improve bid strategies.

We have clients whose web sites gen­er­ate more sales over the phone than they do on the actual site itself. Say they sell com­pli­cated items that peo­ple want to talk through over the phone. We need to be able to tie back exactly which key­words led to sales over the phone, and how much those sales were worth. So, it’s not just attribut­ing a sale to the cor­rect chan­nel, it’s actu­ally deter­min­ing bid­ding based on that data.

Eric Enge: You made ref­er­ence to port­fo­lio man­age­ment, and there is an aspect of that that I’d like to dig into a lit­tle bit, which is the notion that if you are bid­ding on a very high vol­ume key­word it’s really easy to get enough data to make deci­sions about whether that key­word is prof­itable or not.

But, we have the long tail, where the data is scarcer. It’s maybe only a few clicks a day, or maybe it’s a large pay-per-click account that has hun­dreds of thou­sands of key­words that get a few clicks every a week. So, by port­fo­lio man­age­ment, do you mean a strat­egy for look­ing at those key­words in a more holis­tic group fashion?

Bill Mungo­van: Yes, that’s exactly what it is. It’s just an option for us to have two types of bid man­age­ment in the sys­tem. One is the bid rules, which are basi­cally just if-then state­ments. So, if you are get­ting this much rev­enue from a key­word, then you should increase the CPC by a lit­tle bit. But we also have port­fo­lio opti­miza­tion on the other side, which is just another option for marketers.

We found that hav­ing both presents more options to our adver­tis­ers. Now, with respect to the ques­tion of not hav­ing enough data to actu­ally under­stand what’s hap­pen­ing on a key­word by key­word basis for long tail key­words can hap­pen. That’s the biggest fun­da­men­tal prob­lem with port­fo­lios of keywords.

I think the port­fo­lio opti­miza­tion approach does not have enough data, and our tool projects it out based on what lim­ited data we have and what we think may hap­pen in the future. If there is no data, there is only so much we can go on. After a cer­tain point, we just assume that that key­word is just not going to gen­er­ate any clicks. But, that’s one of the prob­lems that we see. We do math­e­mat­i­cal pro­jec­tions for the future based on the lim­ited data that we have.

What I am get­ting at is that our approach to search is the oppo­site of com­pli­cated math­e­mat­i­cal Black Box for­mu­las. We also have that built into the tool, we just don’t believe fun­da­men­tally within Omni­ture that you can click a but­ton and your entire search mar­ket­ing pro­gram will be quickly taken care of. That’s a Black Box approach that we feel has run its course in the market.

We just don’t believe that there is any sin­gle approach to bid man­age­ment or search mar­ket­ing that’s going to work for many dif­fer­ent clients. It speaks to the broader vision in which we view search, which is that we are not an agency, but we have agency ser­vices within Omniture.

Our goal is really to be as trans­par­ent as pos­si­ble to our cus­tomers. Trans­parency is a key issue for us, as we have many clients who have us man­age their search pro­gram for a very short period of time, about three to six months. Then we coach them along the way on how and what we are doing.

We get them up to a cer­tain level of per­for­mance, and then give that over to them in-house. So, not rely­ing on ser­vice rev­enue the way an agency might works to our advan­tage, because we can give full trans­parency to our clients. And that model has been work­ing pretty well for us.

So, to get back to the port­fo­lio ques­tion, the idea of us being able to take care of the whole thing for you is just impos­si­ble in our mind.

Eric Enge: Right. Can you give me a set of things that you are man­ag­ing? There may be ten key­words that are pro­duc­ing great vol­ume, another fifty that are pro­duc­ing mar­ginal vol­ume and then some that pro­duce less than 10% of the vol­ume of the high-volume key­words. You still want to be able to man­age those less than 10% key­words at some level, correct?

Bill Mungo­van: Yes, and that’s an area where you would apply dif­fer­ent rules to the dif­fer­ent types of key­words. One thing we tell clients a lot is there is no faster way to lose money in search mar­ket­ing than to set up the wrong port­fo­lio or to really have poor per­form­ing key­words drag­ging down the aver­age of some of your best per­form­ing keywords.

Sim­i­larly, you wouldn’t nec­es­sar­ily want your high­est vol­ume key­words in the same port­fo­lio as your low­est vol­ume key­words. You may, depend­ing on what the key­words actu­ally are. But you may not, and we want our clients to be fairly care­ful about how they set up the rules in their port­fo­lios if they are, in fact, using that par­tic­u­lar fea­ture. You may actu­ally give more credit to cer­tain key­words because those at higher vol­umes are doing all the heavy lifting.

Eric Enge: Let’s dig a lit­tle bit into the announce­ment you had recently with Scotts MiracleGro.

Bill Mungo­van: In gen­eral, we are doing more and more deals within Omni­ture, both in the Search­Cen­ter busi­ness and in other pieces of our busi­ness that involve mul­ti­ple prod­ucts. And Scotts is a good exam­ple of that, because essen­tially there is only so much you can do if you just think about search engine mar­ket­ing as a silo. So, by bring­ing in data from other sources and using it effec­tively, we opened up a lot of dif­fer­ent options for search cam­paigns. That’s what Scotts is try­ing to do.

We don’t have results for this par­tic­u­lar exam­ple, just because it’s a new announce­ment for us. But, in gen­eral they were hav­ing a hard time under­stand­ing exactly how email mar­ket­ing cam­paigns could be used to remar­ket. And they also want to know how email mar­ket­ing may have impacted or not impacted what hap­pened on their site and what hap­pened in their search mar­ket­ing program.

Scotts was try­ing to take a more holis­tic view of their online busi­ness opti­miza­tion efforts. They made the choice to stop think­ing about email as one silo and search as a sep­a­rate silo. And so, by using Site­Cat­a­lyst as their plat­form, they used Omni­ture Gen­e­sis to pull in the Exact­Tar­get data, and Search­Cen­ter for their search data, and mea­sure it all in one place.

Eric Enge: Right. So inter­ac­tions can be more eas­ily understood.

Bill Mungo­van: Cor­rect. And, on a related note, they had prob­lems with what they called Post-Click Behav­ior, which is basi­cally vis­i­tor engage­ment and what hap­pens on their site once they attract a cus­tomer. They’ve stopped think­ing about email and search as just vis­i­tor acqui­si­tion tools, and started to think about the whole thing holistically.

They can see what hap­pens when some­body clicks on a key­word and comes to their site, includ­ing where go, how much time do they spend on each page and what are they engag­ing with on the site. And they can do the same thing with email as well. Once some­one opens the email and clicks through to their site, what they do and what is most impor­tant to them can be determined.

By using all those prod­ucts in one place, and using Site­Cat­a­lyst under­neath it all, Scotts was able to gain that level of insight. These are rela­tion­ships that we are pulling together these days, because peo­ple want to start to look at online mar­ket­ing more holistically.

Eric Enge: I believe Exact­Tar­get is the email plat­form that Scotts is using, and there is an inte­gra­tion of data between email and the search cam­paign. What are some of the other data sources that can be pulled in and inte­grated in a fash­ion like this?

Bill Mungo­van: Omni­ture Gen­e­sis is the name of the prod­uct that is designed to pull in data from third-party sources. So, Exact­Tar­get is one of many, many email providers we know of.. There is also ad-serving data, which allows dis­play data to also be pulled in.

Another very big cat­e­gory for us is CRM data. By tying actual back­end CRM data to upfront adver­tis­ing, or search engine mar­ket­ing in par­tic­u­lar, you can start to learn a whole lot more about what peo­ple do after the lead has been gen­er­ated. You can also include call cen­ter data.

That can be any­thing that peo­ple do with an SAP or an Ora­cle data­base, any of those enterprise-level sys­tems which may be point-of-sale data, such as data from a sys­tem of kiosks. There are really two types of data: mar­ket­ing data includ­ing online data, such as email and dis­play adver­tis­ing, and offline mar­ket­ing data. So, data from the tele­vi­sion mar­ket­ing or any kind of offline media can also be pulled in depend­ing on how it’s struc­tured in what­ever sys­tem it’s cur­rently in. That’s the one side of the adver­tis­ing data. The other side would be back­end sales data, which is the CRM, Kiosk call cen­ter data and the other enter­prise sys­tems that may live in an SAP or an Ora­cle database.

Eric Enge: Right. And there has to be some pretty inter­est­ing things going on there to pull in CRM or call cen­ter data, which clearly can be mas­sive in size.

Bill Mungo­van: Yes, and we have a prod­uct called Dis­cover OnPremise for when it does get too big. It’s some­thing we got from Omniture’s acqui­si­tion of Visual Sci­ences. For exam­ple, we have a rental car cus­tomer who is try­ing to fig­ure out exactly how many peo­ple book online. Then they’ll go to each indi­vid­ual loca­tion around the coun­try and observe how many peo­ple actu­ally show up to pick up the car they reserved online ver­sus peo­ple who don’t. Then they see what peo­ple actu­ally buy, how far they drive and all other sorts of data like that. As you can imag­ine, it just gets absolutely mas­sive at that point.

So Dis­cover OnPremise is a much more pow­er­ful and robust tool for when inte­gra­tions get well beyond the needs of a stan­dard adver­tiser. But, we do have adver­tis­ers who have mil­lions of key­words in Search­Cen­ter and are tying some of those actions back to the sys­tems that don’t have any­thing to do with what hap­pens on their actual site. So, it starts to get pretty inter­est­ing at that point.

Eric Enge: Are there ways to cre­ate ties into TV adver­tis­ing, print adver­tis­ing and radio advertising?

Bill Mungo­van: Well that’s really the mil­lion dol­lar ques­tion that every adver­tis­ing agency in the world is try­ing to fig­ure out; exactly how does offline data impact online behav­ior and visa versa? And what we pro­pose to peo­ple is to pull that data into Site­Cat­a­lyst, start to fig­ure out your own cor­re­la­tions and, if pos­si­ble, fig­ure out the causal­ity between dif­fer­ent mar­ket­ing programs.

For us, we just pro­vide the repos­i­tory for the data, and then we allow agen­cies and adver­tis­ers to actu­ally start to fig­ure out what is occur­ring on a campaign-by-campaign basis. But yes, you can pull that data into SiteCatalyst.

Eric Enge: What are some of the strate­gies for how you pro­vide the data to Site­Cat­a­lyst? And what kind of data is it that you are pro­vid­ing in some of those more dif­fi­cult scenarios?

Bill Mungo­van: I believe CRM data is the right place to focus the dis­cus­sion because it’s just a lit­tle bit more tan­gi­ble and mea­sur­able. For exam­ple, Omni­ture uses Search­Cen­ter for our own mar­ket­ing efforts in order to get more Omni­ture cus­tomers. A really com­mon sce­nario for us would be to run an online adver­tis­ing pro­gram and then gen­er­ate a lead on a web site. But what actu­ally hap­pens to that lead, at least in our case, is that it then goes to a sales force.

The sales force fol­lows it up, and some per­cent­age of those leads actu­ally turn into cus­tomers. We track it all the way down to how much we spent to acquire that cus­tomer, both online and through our sales team, and then we fig­ure out what we’ve got in return for that. In our case the CRM sys­tem we use is Sales​force​.com. But there are any num­ber of CRM sys­tems from which we can pull the data.

For us, Cost-Per-Lead is a path­way to one very small piece of the full pic­ture, which will actu­ally help us fig­ure out Profit-Per-Click. So, if you are able to fig­ure out how much you spent on all oper­at­ing costs, you can pull that data in through CRM inte­gra­tion, and then actu­ally bid on the key­words that lead to the high­est prof­itabil­ity for your busi­ness.
Those are some of trick­i­est, but most inter­est­ing and most pro­gres­sive features.

Eric Enge: Let’s talk a lit­tle bit about some spe­cific tac­tics. For exam­ple, you know your paid search cam­paign results in phone call orders. And one tac­tic you can imple­ment to make track­ing much more effec­tive is to give every­one that comes to your search from a paid search ad a cus­tom 800 num­ber. This way you can know the results of your paid search cam­paign just based on what num­ber they call into. That’s a tac­tic that is designed to give you much more accu­rate data.

Bill Mungo­van: Yes that makes sense. And another tac­tic that one of our clients is doing is auto­mat­i­cally gen­er­at­ing codes on the site itself. This way the cus­tomer can actu­ally see that code, so each cus­tomer who vis­its the site from a given cam­paign will be iden­ti­fied. And we can actu­ally get it all the way down to the key­word level. We know what key­words they came from that led to a call.

Cus­tomers see a cer­tain code on the site and then make a phone call and either make pur­chase or not. Then we have the call cen­ter actu­ally take that code in from the cus­tomer, so we can record where that cus­tomer came from and what they did on the site. Then we can pull that data back into Site­Cat­a­lyst and make bid­ding deci­sions based on what happened.

Eric Enge: You can also give cus­tomers that walk into a phys­i­cal store a rebate as a part of some pro­mo­tion that the store is hold­ing. Then they col­lect the rebate by going online, fill­ing out a form and plug­ging in the rebate num­ber. Then the web site can check cook­ies to see if the per­son came in from a search cam­paign of some sort.

Bill Mungo­van: Yes. But we wouldn’t nec­es­sar­ily be able to tell what spe­cific key­word they came from. It is still a good exam­ple, but we’ve actu­ally seen the oppo­site hap­pen as well. When peo­ple come online from a spe­cific key­word they come through to a page and have to print out a coupon that con­tains a bar code with infor­ma­tion in it such as the key­word they searched on, bring it into the store and then redeem it in the store.

What we can see there are two things; how many peo­ple print out the coupon and do not go to the store, and how many peo­ple print it out and actu­ally redeem it. So that’s just another way of under­stand­ing what’s dri­ving peo­ple to make offline transactions.

Eric Enge: Exactly. Try to dis­cover every aspect of the inter­ac­tion that you can.

Bill Mungo­van: That’s some­thing really cool that we’ve seen with a retail client. It’s pretty com­pli­cated, but it’s very inter­est­ing at the same time. The client is able to look through Site­Cat­a­lyst as they are run­ning geo-targeted cam­paigns. Again, these are big box retail­ers with stores in many dif­fer­ent loca­tions, and they are run­ning dif­fer­ent ad cam­paigns for the dif­fer­ent geo-locations. And they can see what peo­ple are pur­chas­ing online and, more impor­tantly, what prod­ucts peo­ple are bundling together in a given geo.

So, there might be a video game and CD on sale in the upper Mid­west, and that par­tic­u­lar bundling may be very, very dif­fer­ent from what peo­ple are bundling in Los Ange­les. So what they’ve done is taken all the online data and fig­ured out what prod­ucts peo­ple are bundling online. Then they rearrange the actual place­ments in the store based on what’s hap­pen­ing in that geo. So, when you walk into a store, you would see two prod­ucts next to each other on the shelf based on what peo­ple are doing online in that geography.

Eric Enge: So you basi­cally iso­late the best way to put together bundling based on how peo­ple are behav­ing in dif­fer­ent areas?

Bill Mungo­van: Yes. We fig­ure out what they are buy­ing based on the dig­i­tal shelf and apply that knowl­edge to the actual store and rearrange prod­ucts accordingly.

Eric Enge: What would you rec­om­mend to some­one run­ning a TV campaign?

Bill Mungo­van: It is absolutely crit­i­cal to pull your TV data into the same sys­tem where your online adver­tis­ing data is run­ning. You should at the very least make sure that you are actu­ally mea­sur­ing apples-to-apples in one place. So, it’s not an easy ques­tion to answer in terms of what TV cam­paign yields the high­est pos­si­ble return online. That’s a very com­pli­cated thing, and it will be dif­fer­ent for every customer.

But our advice to the mar­ket on that is to pull the data into the same place and then start to run reports on cor­re­la­tions between media in a given geo and what’s hap­pen­ing online.

Eric Enge: You can also try things like Van­ity URLs, but things like that are very uncertain.

Bill Mungo­van: We have seen a lot of stud­ies that tell us that very few peo­ple actu­ally remem­ber your URL address from the end of your tele­vi­sion or radio ad, and even fewer go on their com­put­ers and actu­ally type it in. For us it’s more inter­est­ing to just let the cam­paigns run sep­a­rately. So, you have your search cam­paign online, a dis­play cam­paign, and then a TV cam­paign. Then pull the data into one place and use ana­lyt­ics to fig­ure out the correlations.

Say you saw a bump on the 21^st of Jan­u­ary, you can find out exactly what media was run­ning in which geo, and then you can start to make cor­re­la­tions between the two. So, I think that those tricks of Van­ity URLs and things like that don’t work in every case.

Eric Enge: Right. Well, I would think that there is a risk of actu­ally low­er­ing the actual return in return for try­ing to fig­ure out how to mea­sure it.

Bill Mungo­van: That’s right.

Eric Enge: Can you out­line how the pric­ing model works for SearchCenter?

Bill Mungo­van: We typ­i­cally charge as a per­cent of ad spent so the more you spend, the lower the per­cent­age. We have cus­tomers in all shapes and sizes. We’ve had clients take it in-house and then they just felt like they really couldn’t han­dle it for a while, and they then requested the addi­tional help of our ser­vices group.

So we man­age it for them for a while, or on an ongo­ing basis, for an addi­tional per­cent of ad spend fee. Then after a while we can give it back again when they are ready. We have some flex­i­bil­ity as part of that offering.

Eric Enge: Can you say any­thing about some other well-known cus­tomers that you have using the ser­vice, and the total spend you have under management?

Bill Mungo­van: Sure. We have 600,000,000 in spend under man­age­ment. One exam­ple of a large cus­tomer that I am allowed to dis­close is Delta Air­lines. They are using an agency called . We have both agen­cies and direct clients using the tool. And we have other retail­ers, like Back​coun​try​.com, using the tool as well.

Eric Enge: Thank you Bill!

Bill Mungo­van: It was good talk­ing to you, thanks a lot!