The Sales Engagement Podcast
The Sales Engagement Podcast

Episode · 2 years ago

Outbound Ops Part 2: Relevance at Scale w/ Ben Salzman and Kyle Williams

ABOUT THIS EPISODE

Max Altschuler hosts Dogpatch Advisors to discuss the future of sales engagement. They discuss the Outbound Ops model, custom code to find relevance at scale, visual prospecting and other concepts from their research and development efforts.

 

Welcome to the sales engagement podcast,this podcast is brought to you by out reached at io the leading salesengagement platform, helping companies, sellers and customer success engagedwith buyers and customers in the modern sales era, check out salesengagementcom for new episodes, resources in the book on salesengagement coming soon now, let's get into today's episode, excited to welcome back dog patch forpart to of their first episode, just had them on recently. If you haven't,please go back and listen to part one. It's pretty awesome we're here to talkwith Kyle and then from dog patch welcome back guys where r you getting your data fromyou're using similar web for this example. But let's talk about datasources, we think about data sources in three main categories. Right, likethere's the data that you buy right so you're, going to talk to vendors,whehther, that's Suminpo, clear bit and others there's data. You wouldoutsource right, so we're seeing this start to crop up more and more, whereteams AE spending manual, effort going and looking up, you know: Does thiscompany have x employees on linked on specifically and they'll? Justoutsource it and th there's collect and collectis either things that you'reasking your teams to do directly or the things that you're automating, thatfirst party data that's really most relevant to your company and often thatin this example for the similar web to so similar web thats fought data, theabstor data that was outsource. In this case we started with having outsourcespull it, and then we eventually automated it. And then the collecteddata was actually the computed attribute between those two data points.Right like the Interlot, the overlap between languagees supported versuslanguages that could be supported based on demand. That was a computedattribute. So we think about by go to the vendors outsource havp. Someoneelse do it where the the throughputter,...

...the CALORI spent, isn't put on yourcellers or your teams that are driving conversations and then collect eitherthings that you're automating, because there's a way you can do it and it'sfirst party data that really matters to your specific brand or it's so valuablethat it's worth putting the teams that your you're paying salaries to go outand collect directly. And you know, I think one of the nextyer asking a really important question like the data pipeline, as we think ofit, is it's probably more important than any individual contributor thatGiveng company has on their team. So a lot of times we're not thinking aboutwhich provider do we pick we're thinking about what are all of thedifferent types of data across companies, people technographics allthe attributes that we could know about any of those categories of companycoming up with sort of backing into? How do we use those attributes? Eithedrive, target or drive perconalization a campaign, and then how do weprioritize and sort of build out a longer term plan for acquiring andnormalizing all that data, and one of the things that we found was when we goin to companies and ask you to what are you doing to source company data? Theybasically say we're going with X, bender an we say why and state of theart response to that in many cases is you know, this is what my DP marketingchose at my last company right, like that's literally, fifty percent of thedata purchase these days are what was my vd marketing doing at the lastcompany and well, that can be a helpful datapoint. The dateo industry moves sofast. It's actually probably wrong for most companies, if you haven't lookedat it recently, so a very large start of a bars, it's a client actually hadus go and run a a large scale test across twenty four vendors globally toreally understand at a statistical, significant level right, we're talkingthousands and thousands of rows of data to say who has which data, where whatare the fill rates for that data, and how would we use that data, and that isprobably the right conversation to have, instead of which vendor should wechoose so most of the companies that were going into are sometimes sourcing.You know five or six different third party vendors, while spinning up largescale, collection efforts and large...

...scile outsource efforts all to bebrought together and sort of inform the self playbook. It's a much morecomplicating process than I think most companies o recognize when they firstgo after this, have you done any reports or insights into us and globaldata providers yeah? So we did both. So we was first for this start of a aneffort to profile the US vender landscape, which we first thought wouldbe pretty clear cut in. It turns out to be quite complicated, there's a lot ofconfusing messaging and some clans thatare dubious at at best, and then wedid it globally and and the global look was really interesting, because wethought we would see a bunch of releas, specific sort of regional vendors whodo really well in particular countries or companies who focus on a particulartype of data. Who would do well there, but not other places, and what we foundwas the folks who are winning today in the market or winning everywhere right.So the best providers tend to have data country to country that is better thanthe vendors who are based in any given country. The vendors who are doing thebest on company data they're also doing the best on people Dhay. So it has alot more to do with building the flywheel and having sort of the DNA tobuild an engineering approach to build a company on the back of data, not sortof thinking O. tactically of you know, we want to go and find a certain numberof companies. It's. How do you build the infrastructure to long termcontinue to collect that data? That makes sense, so I don't know discloseit. What were the findings? Well be careful not to share too much. I don'tthink tho start up that haw US do. This would want us to give all the results.I think what we can say is that the some of the itering to taken away it'swhat one was. There was a cluster of about three or four companies who arein these sort of eighty seven nd. Ninety three percent match rade andwhen we SA Matcretis is this company in your database? What's probably moreconfusing, and most companies really haven't thought deeply about is what isthe fill rate for all the attributes that we care about? So a lot ofcompanies that we're working with we're...

...helping the thing of it? Don't justworry about. Is the company in the database? Do they have the attributesthat you need to do your filter and your targeting its ecetera? What wefound was that was anywhere between about ten percent to ninety percent interms of actually having the attribute so, for example, funding very, verydifficult to pull that obviously crunch base does a pretty good job at when youtalk about Soe of the other global providers, often we're seeing Phil Ransin the sort of twenty five to fifty percent. So you know, when you go outthere and say show me: Every company is raised ten to a hundred million dollars.It's likely you actually have about half the answer, and so it's notactually as cut and dried as people, I think imagine when they start thisanalysis. Ou Know I want to be careful of giving away everything, but therewere some really interesting insights, yeah I'd, say the big things are andand turning this more into like. If someone wanted to do this, what shouldthey do differently than just the? What did we buy at my last company, or oneof the common mistakes we see is the selection of a data? Vender is morepart of a Tam like identifying the size of your market exercise. So then you'reable to tell the board all right, there's a billion dollar opportunitythat we can go capture in business services, because there's this manycompany with business services, great there's, not a different tation in thatvender data, you to go, say something differently or say something relevantbeyond just they have a tag from a vendor. So that importance of sayingwhat are the attributes we actually care about? Not just for who could wetalk to, but why should they want to talk to us and that's sort of the pivot?We're often talking to companies and saying you've already done thisexercise. That says: Who Do we want to talk to, but often there's that missingpiece of? Why should they want to talk to you and that follows that entirespectrum of data you could buy so whether it's technographic data thatare correlated with the types of companies that are ready to talk to youcould be employee size could be presence of certain title could besomething that you go and collect. That tells you they're doing somethinginteresting on their site, but that extra dimension of thinking throughwhat are the attromutes to tell us. Why...

...do certain types of companies want totalk to us today is very different in a lot of these evaluations? So is thiswhat you consider getting into data driven copy? What is that yeah? It's agreat question, so I think that it's a starting point right. You have to havea comprehensive plan and stance on which data is going to be important towhich fill rates that will enable which campaigns, where the density of contaxtthat will have across domands. What technographics can we expect at whatactors Youre Right? So we have sort of answer all those baseline questionswhere Davi derand copy gets really interesting is when a message is actually higher qualitybecause of all thes dated normalization and really taking advantage of anoutbound op function. So in the example that Kyle just laid out from this, youknow language, translation, company, the Openg Hook for a message in thatparticular case is, I noticed you have seventeen a percent of your trafficcoming from France, but you only support English and Spanish. We havesome ideas for how we can help you expand into given country. What wefound was people were responding and saying. Oh, this is really helpfulright. It's not that we're doing Dita difvent copy to save time or because athrepoot is higher, which know it is, but it's because the quality actuallyends up being higher because of the relevant because of the scale andbecause of the automation- and it's really, you know interesting to see thereplies where you actually have people sometimes are playing. I can't tell ifthis was automated or not, but I don't care and that's actually a whole newcategory of outbound that we're really excited, we think of a sort of relevantautomation, and it follows you know very similar trends to what happens aconsumer five years ago. You search for something on Amazon and see a displayad, the next day, you're pissed you're, like Oh, my gosh they're, invading myprivacy. There's all these issues. Now I see a displayout or something likethank you for showing me something relevant right, so that same thing ishappening, an outbound where data is really driving a lot of the quality intheir relevance and people don't actually care as much as even two orthree years ago. As to you know whether...

...there was automation involved, theycare about value and care about whether your vendor can sole a problem for themright. Do you guys use intent data at all surge data, so you mean like searchtrends or keyword data. That type of thing will also look at like. Are theyon GT crowd? You know looking at your company or are they using clurpetreveal right, Yep Yep, so that is an actually it's an interesting seguein othis topic of relevant automation, where we're not so much focused onbeing human, is, like I'm sure, you've seen the examples of you know driftindsegment or these cou, you go to the site and it says you want to see how wecan help outreach and isseys your company name on there. If you did thatfive years ago, you get a ton of letters saying like what are you doing?You're tracking my p address ITD, be this creepy thing and I don't thinkanyone who goes on the site thinks that a human is specifically tracking thatoutreaches on their site and then crafting this message. I don't thinkthat goes into the calculus. It's just more relevant. It's intriguing to sayyou want to see how we can help out reach or the banner changes. So weabsolutely have seen a ton around using intent data. It creates a a better lanefor relevance right because we know there's some type of handrays. It saysthere's some relevance to happen here and it actually reduces some of therequirements of how far sometimes you need to go with personalization,because now we're getting into the tap on the shoulder e and then there'sother ones that are just really straightforward an I think. A lot ofpeople still thin are too Goong to be true right, so anonymous, ips beingsurface as like this person was on your website. It's actually been around forquite a long time. Clearman and others have recently made it accessible right.So the price has gone down to the point where people can identify anonymous,bisiers their site at a relatively low cost, and it does a number of thingsright want Ed. It helps you understand who should we be targeting who'simportant to us? It also does a number of other things across the sale cycleright. So a lot of companies probably end up having business o t website thatare midsale cycle. That's super important to know right. Your midsales,like Youre, scheduling, call you see a...

...bunch of traffic the day before thecall someone cares right or you're closing a deal and all of a suddenpeople start showing up on your side. Guess what it's financed, wonderinglike who are we contracting with there's so many signals that you canpull out to better understand, what's actually happening on the other sideand again it's much more accessible than it was even a few years ago Iwould say the one trend we've noticed to be interesting to see what you seeon this Max. Is, you know around saying you don't have to prove that you'rehuman, you can just be relevant? Is there still a creepy line that existsso calling out and saying hey, we analyze e Ip traffic on our site andsomeone from your company was on our side. Are you interested probablydoesn't pass that bar and it's what Ben and I call the Shakespeare effect,which is you're going to dinner with someone you see they've likedShakespeare on facebook. You don't sit down and say I saw you like Shakespearon facebook. Let's talk about that because it could be a number of reasons.Why, right I did a paper in tenth grade and I had to click the box or why areyou on my facebook, looking at what I like and using that for topics what wewould say to do in status, use that intent data to say something relevant,so you should just quote: Act Threeline, two of Athello and if they really likeShakespeare, then they'll say oh, my gosh. I Love Shakespeare and, if not,then you said something kind of smart and it passes by harmlessly it's thatconcept of using that intent data to be relevant, but you don't necessarilyhave to call everything out. I like that. I like that a lot. So let's talkabout that, some more so you're talking about you were talking about earlier.The two by two Matrix, the goodness of fit for vendors, was Tho, goodness Ofit for prospects. How does that work into these plans? Yeah so and oftenthat's even a a selection. So we have to talk about the difference betweenyour targeting data and your messaging data, and what often gets mixed up istreating those the same so usually ou end up locked in this dicotomy of we'regoing to send scale messaging, which is we're going to make a list and they'regoing to write the message that matches everyone on that list or we're going towrite, superrelevant messaging and so...

...we're going to go one by one and writethe message for each person going to match the message to each person and wewould say, there's a combination of the two. So you get the list to a certainsize and then you adapt the message within that list. And often thosepoints are going to be. Who are actually the people who care about usand why so that you know that Matrix is basically saying on the one side. WhoDo we really want to talk to and why right like they have a certain amountof funding they're in a certain business model. And then what are allthe reasons that people want or don't want to talk to us right. So if you'rea startup, there's a certain class of company that doesn't like talking tostart ups and there's certain class company, that really does like to startto talk to startups. So there's, probably thirty different attributesthat are a plus or a minus for the messager trying to send, and for someof those you would simply not send it right. Now we're not going to focus onthat particular part of the market right now we're going to win where wecan can compete and then, within a certain amount. It's. How do you adaptyour message to talk through? You Know Ben, and I talk about this example ofwhen Wewere at Google and this wis before the data providers. So I'dwritten this little Bot that would ping email servers because email serverswill say: Hey I'm alive, I'm here, but every span provider, email server has adifferent flavor of how they give that what's called a hello command. Soprofiled, though so we could know, all right are using Sisco Ironport, whichis really expensive and the heavy it decision or using something open source,and you could totally change the message based on Cisco Ironport. I'mgoing to talk about how Google has five hundred security engineers and were themost secure platform for this, in this case towling Gmail. But if you havesomething open source, I'm going to talk about cost of ownership and howyou don't have to deal with Maitenance, and it's that difference of like tuningthe message versus deciding, maybe who you don't even go after yeah and Ithink iy'll give you another story. You know from our last start: U Bo Kylesfirst day at a mobile market, compat that we both worked at. He asked you.How do we identify the week competitors and, I said Yeah- I wish that that wasout there right. This is literally...

...predatinize right o. He had nothing toknow what people had tech and Stalland. There was a bunch of we think of hislegacy providers out there who are doing relatively uninteresting thingsand Kyle said. Okay, I'm going to in all of my glory, write some code tosniff my own WIFI traffic, download, the top thousand haps in the APSTOREand see what the payload contains. So we were able to all th sudden identifyhere's. What sdks are installed in all these different APPS and it was thelowest hanging fruit right, you find the weakest competitor and you startyeretarting those people straight away. So what's interesting is the explosionof raw data makes it so not every company has access to that and itcreates this downward pressure on relevance because everyone' sendingthat same thing right so someone's I notice you have xteck installed and Iwant to sell you my product, that's very focused on what you want for yourtarget customer right. What is your ICP that doesn't really speak to whetherthey want to talk to you. So, what's much more powerful again is going fromobservation, which is you I saw you have Xtech Ansol to inside, so I sawyou have xteck installed and you've hired a number of new employees a areexpanding into this new market. Here's how we can help to solve a problem,that's created when all three of those things are present again very easy foran SR to go figure that out in an hour for a single company very hard to dothat across your entire targeting data set or your ICP, without the rightfunction in place to go and sort of normalize that data to be used. I love how you guys have merged, andyou know, engineering and sales to really like masterrlly put togetherthis outbound ops. I don't know if it's a new role or position or industry orwhatever it is, but it's certainly necessary. You know it sailes, ops,sales, op sooped up for for outbound for pipeline building right totallyyeah, I mean look. The sales opsfunction has been important for along time. They never really touched outbound because it was such a humandriven and manual thing. They didn't really need to, so we do think of it asa new function. We also think of it as a new model right so th. The idea ofthe sort of predictabale revenue model...

...and Aron Ross inventto this in twothousand and two he's done a lot for the industry. The world simply changedaround that model right, it's so from a time when you K, ow specialization wasbreaking out quota carrying raps into prospecting, raps that samespecializations happening in outbound right, so it's been having a specialsations at having the economy for a hundred years. In this case, there'sall this raw data, this sort of explosion of structure data that exista power outbound, so it is a new function and a new model, and- and youknow, as we've said- it's not that you know it's data for Dayta Sake. A lot ofthe relevancs as coming out of these messages is driven from the Rawdeta,not from the human. So for a lot of our customers, woul talk about sort ofmodeling out their head cout or their growth projections. We would, in almostall cases, recommend a company goes and spends seventy to a hundred thousanddollars on more data than another single individual who you know, maybefully loadd in the area is going to be close to that hundred Kmart, there's somuch more leverage that comes out of the the data than there is from sort ofl one by one or linear effort yeah. I agree and really love what you guys aredoing. So you work with companies like clear bit or wee allowed to get in asotof segment. I think there's another one. Your guys are in some of them haveto blieve at least one of those but yeah G. Okay. Well, you guys areworking with a lot of the fastest growing companies in Suckom valley andmost forward tinking companies workg on out Boun ops were excited to get you onfuture episodes of the show to talk more about kind of the future ofoutbound, and you know some of your tool kits and stuff like that, butreally appreciate, avinoon and taking time out of your day to talk to ouraudience and looking forward to working with you more. Where can people findyou right now and if they want to learn more yeah men, good o website, we putup a decent amount of content and we sometimes speak at events. But you knowjust come talk to us right. We love talking to companies about, what'sgoing on in the world of sales playbooks and out and out bound andreally focusd that you sort of mentione on the very highest end companies rightlike we don't tend to be the company who comes out with a report about whatar a hundred sales leader saying across...

...the industry. If it's the middle of themarket, we don't care, we're really focused on whate. Are the very highestend companies doing to push the conversation forward, and we lovetalking do about that. So you know come find us on twitter, linked in whereverand Yeah Max, really appreciate you having a song were excited to be onagain and congrats on everything you guys are building at sales hacker andoutreach. It's Awso thank so much backs. Thank you, yeah most Siana futureepisode and that's all fored. Today. This was another episode of the salesengagement. PODCAST join us at sales, engagementcom for new episodes,resources and the book on sales. Engagement coming soon to get the mostout of your sales engagement strategy, make sure to check out out reach te IO.The leading sales engagement platform see you on the next episode.

In-Stream Audio Search

NEW

Search across all episodes within this podcast

Episodes (308)