Latest / State of the AI Union / Tamar Gill: The 80+ Step GTM Data Validation Process
Transcript
- 0:01Hello, everybody, and welcome to this week's episode of the State
- 0:04of the AI Union. I'm your host, Laura, and join
- 0:07with me today is Tamar Gill, CEO of Ecor.
- 0:10I'm going to let Tamar introduce herself and also what Ecor does.
- 0:13Thank you so much for having me, Laura.
- 0:15I'm Tamar Gill, CEO and Founder at Ecor and I fixed that data in
- 0:202020. I started Ecor with two main
- 0:23principles in mind. One was to fixed bad data issues
- 0:28for go to market teams and two was to give job opportunities to
- 0:32economically disadvantaged communities, specifically stay
- 0:35home parents who want to contribute to their family
- 0:39income and grow the same time while being around their
- 0:43children. And today I'm proud that six
- 0:46years of bootstrapping, I ran E core with main principal that
- 0:52people first that only work with the clients where we bring the
- 0:55huge value. And the main thing is that we
- 1:02are a real proof that you can actually build something real
- 1:08honest without outside funding, without hype and without
- 1:14anything that's happening today. I love that and as a parent of
- 1:18four children, I also feel that, you know, providing parents with
- 1:22the opportunity to work inside their home is super important.
- 1:25So thank you so much for doing that.
- 1:27We'll come around back to that question later, but let's talk
- 1:31about go to market and data. And I was managing Rev OPS teams
- 1:39for a long time. And so I always notice this data
- 1:43quality problem at the time of out outreach and outbound when
- 1:47emails start start bouncing and then people complain, oh, you
- 1:50know, my data is bad. You know, the campaign isn't
- 1:53working. So we noticed that when the when
- 1:58the effects are are are obvious, right?
- 2:01But how bad is this data problem actually?
- 2:05Yeah, that's a very good question.
- 2:08Currently, R.E.M. Data is only 26% actionable. 26.
- 2:1526% actionable. So that's how big the data
- 2:19problem is today. You know, and e-mail addresses
- 2:22and that those are little things when you notice that you know,
- 2:25like emails are bouncing. By that point, damage has
- 2:28already spread entire across entire go to market motion 2074%
- 2:36of non actionable data that breaks more than just campaigns
- 2:41breaks your pipeline, it breaks your peephole and it breaks your
- 2:47predictions because everything depends on the data.
- 2:51Every company runs on data right?
- 2:54But you know what's the scariest part that most teams don't see
- 2:59how bad their data is because it looks complete and when they
- 3:03look there CRM, it looks healthy on the surface right?
- 3:06But underneath sales is working wrong accounts, marketing is
- 3:11targeting wrong people and Rev OPS spend massive amount of time
- 3:17and you were in the Rev OPS and you know that trying to fix the
- 3:20bad data issues. That's right.
- 3:22That's right. And so we are already applying a
- 3:28lot of AI on top of the on top of the data that exists and and
- 3:32the data as you said is only 26% actionable.
- 3:35So we're doing a lot of outbound sequencing, you know, lead
- 3:38scoring, etcetera and all those things.
- 3:42Tell us what we should be doing with AI and the data.
- 3:46OK, Oh, lead scoring, I think AI can help with that only when
- 3:53your data is already clean. Now you apply your, you know, AI
- 3:57to a dirty data, what kind of scoring you're going to get,
- 4:00you're going to score manager. That might not be relevant and
- 4:04you're going after VPS or directors, right?
- 4:07AI can just score your manager as your decision maker and many
- 4:13other things. So the first thing I always
- 4:16recommend everyone I talk to is before you apply AI to anything,
- 4:21check what's in there. Check your data, run the simple
- 4:26validation test right. Take a random sample from your
- 4:30CRM and see first of all, check employment status.
- 4:35Are these people still at the company?
- 4:37Not just you know if the e-mail is valid or not?
- 4:40Yes, employment status and check are these contacts mapped to the
- 4:43correct accounts? Because I'll tell you a fun
- 4:46story about it. They're not mapped to the
- 4:48correct accounts. And after that I keep the e-mail
- 4:52is followed, right. That's the three main things you
- 4:54want to do quickly to fix that problem and I'm here to help for
- 4:59that. We actually run free validate,
- 5:01free assessment, CRM validation, you know, assessment or the
- 5:07organization so we can show the current state and quick, quick
- 5:12story about like why it is important, right?
- 5:15So my HR director receives an e-mail that starts with as a
- 5:22leader of, you know, HR at the company she left three years
- 5:26ago. No.
- 5:28We are any she received this e-mail on her current job e-mail
- 5:32address. Yes, yes.
- 5:34So and then, you know, blah blah blah, whatever e-mail said.
- 5:38And the reason for that is exactly because when people
- 5:43enrich data, they don't check if the accounts and e-mail are
- 5:47mapped correctly, yes. So that's why like checking and
- 5:50validation is crucial. Yeah.
- 5:53And that's really interesting also because you know, sales is
- 5:56about trust, right? And if, if I can't trust you to
- 5:58even know which company I'm working at, how can I trust you
- 6:02to have a conversation? So you know, the the starting
- 6:05point of where we connect, I think it's like super important.
- 6:08Absolutely, absolutely. It breaks trust right away.
- 6:11I would never, never work with them.
- 6:12I would never go and work with them.
- 6:13That's where I want to reach out and say, hey, where do you get
- 6:17your data? You need help.
- 6:18You need a better data provider or better process, right?
- 6:22Because they go and reach this and my God, this is cool.
- 6:26We had reached, we got so many emails.
- 6:29It's perfect. Well, no, it's not.
- 6:31It's missing the crucial. Part and tell us a little bit
- 6:34more about how that validation can happen with AI.
- 6:40You know the one that you mentioned about about validating
- 6:43a random sample set? I did not say you can validate
- 6:46it with AI. OK OK, so is the validation
- 6:49manual then? You can use AI to help you
- 6:54validate, but AI is really good on like contextual reasoning.
- 6:58For example, AI can determine if the company is still active,
- 7:02inactive, acquired, merged or if there is relationship between
- 7:07two companies to compare or if the record is duplicate.
- 7:12Like if two you know 2 records are duplicate or not.
- 7:16As well as if you know similar name companies are belong to the
- 7:21same organization or totally different organizations.
- 7:25Right, I can help you with that. But the validation is pretty
- 7:30complex steps like in our case we have 80 plus step process
- 7:39that validates and reaches your data only after validation.
- 7:44Got it, got it. 80 plus steps. Tell us a little bit more about
- 7:47what those 80 steps and until actually.
- 7:49My God, yes. So it's a very, we build this 80
- 7:54plus steps based on the manual and human work, human research
- 7:58that we were doing right. So we are very focused on
- 8:02quality data. We advise everyone to validate
- 8:06before and reach and based on all our experience, which in my
- 8:11case, I'm 14 years already in data space and six years in my
- 8:15company. We learned a lot on what all is
- 8:19required to give you the perfect data and that requires 80 steps,
- 8:26which I'm not going to list all 80 steps right now, but it
- 8:31definitely goes through. Is this the right person?
- 8:35Is this person still at the company?
- 8:37And if this person is still at the company now, is the title
- 8:41still valid? If the title is still valid, you
- 8:43know, or if the title changed, is the job function relevant
- 8:48right now for my you know use case?
- 8:50Or is the seniority level still relevant?
- 8:53So it checks every single thing based on also your ideal
- 8:57customer profile as well as validity of the record itself.
- 9:02Got it, got it. So you're saying that that the
- 9:04enrichment process has has those 80 steps, but we only do that
- 9:09after we validate that the data is correct.
- 9:12And earlier you had said that the validation process in, you
- 9:16know, just sort of manually verifying that these that that
- 9:20the data records are, are good. Like did I hear you correctly
- 9:23that you know that is usually a manual process just to?
- 9:27Double check that not for us now.
- 9:29This 80 plus step includes validation.
- 9:31Includes validation. OK, got.
- 9:33It so yeah, first like 4050 steps, validation steps, OK and
- 9:39after that starts enrichment. Got it, Got it.
- 9:42And in in this 40 plus steps that that are using validity, is
- 9:47there a component to how AI is helping you do that?
- 9:50Of course. Tell us a little bit more about
- 9:52how you do that. Of course, there are multiple.
- 9:54Definitely we have incorporated AI in our steps.
- 9:58There are probably more than 30 steps where we have AI and
- 10:02mostly it is helping us determine the relationships.
- 10:07So when we give it like two different things, we wanted to
- 10:10compare and tell us in this this, you know, matching, for
- 10:14example, we have current company name and previous company name
- 10:17and we want to understand is this a change or is this just,
- 10:22you know, some kind of merger or rebrand or anything similar?
- 10:26So it's very good in helping in those situations.
- 10:31Again, when we have situation where two data points like New
- 10:35York and New York, like it doesn't match, right?
- 10:37Like is this an error or is it just formatting issue, right?
- 10:42So there are many steps that AI is helping us and it making it,
- 10:46you know, much faster. And definitely we're training
- 10:50it. We're still training it.
- 10:51We're still teaching it because you cannot blindly trust AI
- 10:55currently. But we've, you know, we used all
- 10:58our, you know, human research experience to train the AI for
- 11:04that. Got it so in most Rev OPS
- 11:06organizations, including my own, like we source data from
- 11:09multiple data providers and they all have their own validation
- 11:16systems, right And and a lot of them tell us, you know, our data
- 11:19is validated. It's, you know, some of them
- 11:21have like platinum level validation or diamond level
- 11:24validation, that kind of thing. What would you say is changing
- 11:28in the way that you're doing it today versus maybe how these
- 11:31other providers have been doing it for the last 15, you know, 20
- 11:34years, so. One of the biggest difference is
- 11:37that all traditional data providers run it against their
- 11:42database. So they have their database
- 11:45OF600300 whatever number million records they have.
- 11:50So they runs against the database and it returns matches,
- 11:54right? It either matches or doesn't
- 11:56match. And then based on matches, in
- 11:58most cases it's determined if the person is, you know, while
- 12:02it's still at the company or not.
- 12:04And in many cases they don't even tell you if the person left
- 12:08the company, right, it will return the new data.
- 12:10And now you have to figure out what to do with it next.
- 12:14Now the difference now in our case is that we don't trust
- 12:20anyone database. We do have our own 500 million
- 12:24records but and we refresh it quarterly, but changes happen
- 12:28daily. So we don't trust it every time
- 12:31we run the validation and we match against our database.
- 12:35That's why it's 80s plus steps. We then run it in real time to
- 12:39check you know if even we're checking if our database data
- 12:43was correct. So that's the difference.
- 12:46It in real time to understand if this data is actually valid.
- 12:51Yeah. And and I also hear you saying
- 12:53that you are, you know, relying on like multiple other sources
- 12:56as well to make sure and, and checking all of these different
- 12:59versions to make sure that your data is constantly refreshed and
- 13:02updated. Yes, definitely we, we refresh
- 13:05our data quarterly. We also partner with many other
- 13:08organizations, other data providers to increase our
- 13:13coverage, but anything we pull from anywhere goes to
- 13:17validation. None of them goes to our clients
- 13:20straight. It has to run this validation
- 13:23steps so we can reach that 80 plus percent.
- 13:27I mean I'm sorry, T plus percent match rate and accuracy.
- 13:32One quick note, currently when on the market when the providers
- 13:37run the data against their database, you will see about 30
- 13:40to 40% match and with us it's above 90% because we multiple
- 13:46database plus real time. Yeah, You have a do you have a
- 13:49best practice on how organizations should be thinking
- 13:53about their go to market data and the refresh rates of it?
- 13:57I know for example, from my internal customers, they're
- 14:00always thinking that what they want is I want all the data all
- 14:04the time, you know, like every single second.
- 14:07It just has to be updated in real time.
- 14:09What's your best practice here on using and updating good
- 14:13market data it? Depends on the size of your CRM,
- 14:16right? When you're enterprise and you
- 14:18have millions of records in your CRM, you cannot update it like
- 14:24all the time, right? For that, we recommend at least
- 14:28quarterly refresh. If you really want to do it more
- 14:33like, especially I'm talking about when you have 40 million
- 14:35records in your CRM, right, quarterly is a good way to go
- 14:39because you know the first initial run will eliminate 40%
- 14:44of the data already because that 40% people are no longer at the
- 14:48companies anymore. The rest will get updated and
- 14:51enriched and then quarterly you will get an update on what
- 14:55changed specifically about the changes.
- 14:58Now if you have a smaller size data in your CRM, you can we
- 15:04actually do run daily monitoring for them.
- 15:07Weekly or monthly depends. If you really have high target
- 15:11contacts, then it's going to be daily, right?
- 15:13You don't need to Daily Monitor millions of records, right.
- 15:17Most common use case with us is monthly for a smaller scale and
- 15:22for the larger one it's quarterly.
- 15:26And again, when I say we validate their data, their data
- 15:31goes in real time in that moment and gets validated.
- 15:36Right, right. And so there is a way to
- 15:38maintain it without, you know, like Rev Up spending 80% of
- 15:43their time trying to manually fix it.
- 15:45Yeah. So I'm hearing like a quarterly
- 15:47refresh is probably you know. Quarterly refresh for a really
- 15:50large data is good. If you're talking about smaller
- 15:53one, anything like under 1,000,000 monthly refresh?
- 15:57Got it, Got it. What are you seeing in terms of
- 16:00the level of difficulty of getting accurate data today?
- 16:05Is it actually becoming more difficult or is it becoming
- 16:09easier with the different tools that we're given the?
- 16:12Challenge is because we take so many tools that plug them all in
- 16:17and it gives us conflicting data, we are lost and we create
- 16:21more chaos. Again, More data providers and
- 16:23more tools does not mean you have better data, right?
- 16:28For better data, you need maybe one or two quality data
- 16:32providers, right? That will follow your rules and
- 16:36your formatting rules, because you will get from many providers
- 16:39all different formats. And then what happens?
- 16:42Like how can you segment with all these different formats in
- 16:46your CRM? It's almost impossible.
- 16:50And then routing and everything breaks because of that, right?
- 16:53So many providers doesn't mean you have good, good data.
- 16:57Maintaining and cleaning your data and getting to start with,
- 17:02right, getting only the data you need.
- 17:05You don't need most of the data you get.
- 17:07You only need like a portion of it.
- 17:10So get what you need. And when you have the amount
- 17:13that you really need, then it's easier to maintain it.
- 17:17And then you don't need so many providers to give you data,
- 17:22right? Even you know if you need
- 17:24millions, you don't need too many providers because it
- 17:27creates chaos. Right, right.
- 17:30With, with your, with your, you know, 30 step validation in, in
- 17:35this 80 step enrichment, you're, you're leveraging AI to help you
- 17:40with that validation. And, and today we've got a lot
- 17:42of new tools that or I guess AI allows us to do a lot more with
- 17:49data enrichment and validation. What are some things that we can
- 17:55do a lot better today with with AI and why do you think more
- 18:00teams are not taking advantage of those things?
- 18:03There is a Little Miss understanding on like how AI can
- 18:08help in that, right? Because some people think that
- 18:11you just take AI and you just create like one agent and it can
- 18:15fix everything. You know, it will fix your flow,
- 18:17it will fix your data. That's not true.
- 18:20AI cannot magically fix your data, but to take advantage of
- 18:24it is exactly how we took advantage of it, right?
- 18:27We're using it where we are seeing that it's working the
- 18:30best. Those few cases that I mentioned
- 18:35like relationship detection, duplicate detection and so on.
- 18:38So that's the best way you need to find where you're using it
- 18:42best and what type of work you're trying to do, what type
- 18:46of validation or image. But AI cannot tell you if
- 18:49someones e-mail is valid or not. That's not a is job right?
- 18:54It cannot tell you or if you don't give it enough data, it
- 18:58will not give you especially at scale if this person is still
- 19:02employed or not. AI needs to have some data
- 19:04right? It needs to have current
- 19:06company, company from your CRM and current company to
- 19:10understand if this person is still there or not.
- 19:13So you just need to feed it the proper information for it to
- 19:17return good results. So if I'm hearing you correct
- 19:21correctly, maybe what you're saying is that what we've been
- 19:26trying to use AI for is maybe not the right application of it
- 19:31on on the data side. And you know what we're trying
- 19:35to get it to do magic that it wasn't designed to do.
- 19:39And, and as a result, it can actually become worse.
- 19:42It can actually become worse for you.
- 19:44Whereas what we should be doing is using AI for the things that
- 19:48it it can do, for example, deduping and you know, sort of
- 19:53like maybe running refresh rates, things like that.
- 19:57And and that's probably where you would recommend people
- 19:59start, is that? Yeah.
- 20:00Yeah, yeah, definitely. That's where it fits in best.
- 20:04Refresh, maybe not, you know, not refresh, but you have to
- 20:08have a process, right? You cannot just put AI on the
- 20:10CRM and tell hey check my data that that's not going to work.
- 20:14Right, right. Right.
- 20:15So you need to have some kind of flow workflow built for it to
- 20:19check, right? But for that, you still need
- 20:22some other, not necessarily AI, but other steps in place.
- 20:27Yeah. In terms of what you're asking
- 20:30for it, yeah. OK, cool.
- 20:33So for for people that are listening or let's say buyers of
- 20:38data, buyers of data, if they wanted to do something different
- 20:43about their, the way that their go to market data is being used
- 20:48or processed, what would be a couple of things that you would
- 20:51recommend to them today? One thing I recommend everyone
- 20:55is first diagnose your data. That's for what you have
- 20:58diagnose it. The verify the, you know, status
- 21:03of the person, you know, employment status, and then
- 21:06everything else that's step #1 #2 when you buy new contacts.
- 21:13I'm going to tell you a little bit about how that works, right?
- 21:15So you have target accounts and you want to buy ideal personas
- 21:22for these target accounts. So what you do is you upload
- 21:26these accounts in Zoom info Apollo, whatever provider you
- 21:29have and you want to get marketing directors, for
- 21:33example, directors of marketing sales and IT.
- 21:38So these are your three ideal job functions and you need job
- 21:42level needs to be marketing. When you pull that list, it
- 21:46gives you a lot of irrelevant titles because none of the
- 21:49filters are perfect. You might end up up with finance
- 21:53manager. Did you request finance manager?
- 21:55No, you might end up with people that left already the company
- 22:01and you will also end up with people that work in similar name
- 22:06companies and not necessarily those companies, right.
- 22:09So even when you buy this new data, I recommend to validate it
- 22:14first before you push it in your CRM or outreach.
- 22:18Because what we're seeing is when we do similar exercise for
- 22:23our customers, we remove over 50% of irrelevant data and
- 22:28that's a big number. Now imagine how much the data
- 22:32ends up in your CRM and then you put them all in the campaign.
- 22:37That should be perfect for marketing directors, but you're
- 22:40reaching finance managers. So diagnose first, like find,
- 22:46find the gaps in that diagnosis. What are some things that you
- 22:50recommend as steps? Well, first definite step is,
- 22:53are these people still working at this company that you have in
- 22:56the CRF? That's number one.
- 22:57How? Would you?
- 22:58How would we do that? What's the best way to do?
- 23:00That well, for one of the things again, most of the organizations
- 23:06internally probably will, should take like 50 records and check
- 23:10it manually. Because if I tell you to go and
- 23:13try some tool, it's going to that's not going to return it
- 23:16because only like 30 to 40% will be matched.
- 23:18So you're not going to get a full, full picture.
- 23:21If you want to do a small scale, do it yourself.
- 23:23If you want us to test a larger scale for you for no cost, we
- 23:28can do it for you because that's what our 80 plus step flow does,
- 23:32right? They can come to us and no
- 23:36charge will will test it for them.
- 23:39So take a sample set. Do you know a short validation
- 23:44just to see, hey, what, what are the, what are the gaps?
- 23:47And then as you're buying data to fill in the gaps, make sure
- 23:49that you validate that before pushing it into your serum
- 23:53because otherwise you'll get plenty of undesired records
- 23:56that's going to mess up all your your CRM hygiene and stuff like
- 23:59that. Yeah, yeah, definitely.
- 24:01Because without validation, if you put it straight to
- 24:03enrichment and not validate, you will end up in this situation
- 24:07where wrong accounts map to wrong people and receive send
- 24:12out the emails just like my HR director received.
- 24:15Right, right. And then you're just wasting
- 24:17your money and resources on person that's not relevant for
- 24:21you anymore, because that account was for you relevant.
- 24:24They would want that that person is now might not be relevant,
- 24:27right. So first you need to know if
- 24:29this person is still at the company or not.
- 24:31If they're not, you need to archive the person, right?
- 24:34Yeah, if you definitely want to go after this person, then get
- 24:38fully new information, where they went, their new company
- 24:42information, and then, you know, add it as a new record in your
- 24:46CRM. Otherwise archive this and
- 24:49whoever is still at the company, then do the additional checks on
- 24:53other data points, right, like function level, title, so on.
- 24:58Once that's done, then whatever data points you're missing, you
- 25:01can use additional data providers.
- 25:04That's where you can apply data providers to fill in that gap.
- 25:08Yeah. So, Tamar, this process used to
- 25:10take us like between three and six months and we would do them,
- 25:15you know, like every like every year or so.
- 25:18And it would take a long time to to do all, you know, all this
- 25:20validation because we had so many steps.
- 25:22What do you think are the more appropriate timelines today?
- 25:24No, no, no, no. I mean, it depends on if you're
- 25:28running a really like reputable organization and you want your
- 25:33outreach to be also reputable, then they they'll delay this
- 25:38quarterly. Again, this process used to
- 25:42also, you know, it was a long validation process and ritual,
- 25:46but now it's not anymore, right? Because now there are companies
- 25:50like us. Well, that's that's not correct
- 25:54because nobody has as comprehensive like the tool that
- 25:59we built because, you know, we know what's out there.
- 26:04But you can partner with someone that will do it for you and you
- 26:08will find the huge savings. That's one monetary savings.
- 26:14The second is time and your reputation first of all.
- 26:20Finally, all can you know your team can start trusting the data
- 26:24you have. Yeah.
- 26:25So would you say like if we were doing a refresh in data like a
- 26:28project like that sort of the first steps would be sort of
- 26:31more like in weeks like would start to see value in terms in
- 26:35in weeks like? Size.
- 26:36What size? Let's say you know a company
- 26:39that has 500 million in revenue. I mean what size of CRM data 5?
- 26:47100 million records, let's say. 500 million records that would
- 26:51be a database like us right yes. So we are the database that has
- 26:55500 million records and to maintain this database, you run
- 26:59it, we'll refresh it quarterly, OK, quite whatever O
- 27:03organization, you know, like it's just data providers that
- 27:06will have that many records. Otherwise what I've seen so far
- 27:10is the highest numbers where 40 million and then 80 million.
- 27:17And I'm talking about really large organizations.
- 27:21And the reason for that is that they also archive from time to
- 27:25time they archive the data. If they have 500 million, that
- 27:29means they've never archived the data.
- 27:32Yeah. So for them you know, first step
- 27:36will be remove all the and archive and that will be just
- 27:40what and then it will be left with probably 200 million
- 27:43because they've never done the validation this 200 million and
- 27:47this 200 million will be refreshed quarterly, yes.
- 27:50Yeah, OK, OK, good. Awesome.
- 27:54Tamara, let's learn a little bit more about you.
- 27:55So I know you're out of Pennsylvania and you started
- 27:59this company out of there. What what is it like?
- 28:02What is it like starting a company, you know, in a place
- 28:05that's not like one of the hubs, like Silicon Valley or New York?
- 28:08A very good question. And as you can see from my
- 28:12accent, I'm not originally from the USI came from a Republic of
- 28:17Georgia that's part of former Soviet Union.
- 28:21And my background was in nonprofit organizations where we
- 28:24were actually making a difference in people's lives on
- 28:28a daily basis. So, and I was on the executive
- 28:32team there and from, you know, Georgia, I came to, I relocated
- 28:37to the US and became stay home mom and and I became stay home
- 28:40mom. I enjoyed it for a little bit,
- 28:42but then I realized, you know, I need to do more.
- 28:44I need to work and I want to continue growing.
- 28:49So I started looking for something that would help me
- 28:53stay at home with my children and the same time grow.
- 28:59So I found this company that was in Silicon Valley.
- 29:03I joined them and I worked with them for seven years.
- 29:07I started on their lowest role and ended up on the executive
- 29:11team because of my performance and how appreciative I was of
- 29:17the the opportunity, right? Because, you know, I got it.
- 29:20I'm at home, I'm with my children, I have a job where I
- 29:23can grow. And it was really beautiful.
- 29:26And then when I started Ecor, I had the similar sense in mind
- 29:30that there are so many people out there.
- 29:32I mean, you don't have to be immigrant.
- 29:33You can be in your own country, whatever, but you're at home.
- 29:36You want to be at home with your children.
- 29:38You don't want to commute, but you also want to earn some
- 29:43income as well as grow. So that's kind of how we started
- 29:46E core and we're we built it, you know, slowly to where we are
- 29:51today. Being in Pittsburgh definitely
- 29:54is a challenge because there is not many business related
- 30:00things. Yes, lately we have a lot of
- 30:02tech companies starting in, in Pittsburgh.
- 30:05So I've been making some connections here as well.
- 30:08But one of the main, you know, things because most of my
- 30:11clients are either in New York or San Francisco or I do travel
- 30:16a lot to meet them where they at and of course to, to connect
- 30:22with others because as you just said, Silicon Valley is where
- 30:28tech company needs to be. Yeah, yeah.
- 30:32What a great story, Tamar, You know, to be an immigrant here
- 30:35and also, you know, starting your business with, with your,
- 30:40with your kids at home. I think that's, that's really
- 30:42inspiring. It's inspiring for me.
- 30:44What advice might you give, you know, some younger listeners
- 30:48that are thinking about starting their own businesses and
- 30:50thinking, oh, maybe it's not the right time because of, you know,
- 30:53life changes and family commitments.
- 30:55What what? What kind of advice would you
- 30:56give for? That there is never a right
- 30:58time. That's number one.
- 31:01It's for me, it was, you know, I felt like, you know, jumping
- 31:05without a parachute when I when I started equal.
- 31:08So I had two options, go back to corporate or start something on
- 31:11my own. And then I thought, if I can
- 31:17contribute so much to someone else's business, maybe I can try
- 31:20my own. Exactly.
- 31:22So that is what got me pushed, gave me a little more push to go
- 31:27on my own and start on my own. Yes, it's a huge risk.
- 31:31Yes, we have statistics where a huge number of companies fail in
- 31:36their first, you know, year or two right.
- 31:39But one thing is the key. If you start don't have the
- 31:44mindset oh, if this is not going to work, I'm going to do this
- 31:49No, your mindset should be this is going to work because I'm
- 31:53going to make it work and then you work you write well, all you
- 31:57need to do for it to work and then follow step by step Yes,
- 32:01it's a 24/7 when you run your own company, it is hard, but
- 32:05it's not like 5:00. I can close my computer and be
- 32:09done. Absolutely not because my
- 32:11company is like, you know, part of my family, like sometimes
- 32:16it's it's personal because it's my own right.
- 32:18So you care too much to you want to respond to your client, you
- 32:23know, like within minutes, you don't want to keep anyone
- 32:26waiting. You want to solve every issue
- 32:29right away. So it's going to be a very hard,
- 32:32but you can do it if you really put your mindset to it.
- 32:37And that's one thing I always tell my children, you can
- 32:39achieve whatever you want. You need to think about it right
- 32:43way. Yeah, that's a really, really
- 32:45good advice and and super inspiring.
- 32:47And I love how you said, you know, it's all or nothing, like
- 32:50go all in. Like, you know, like make this,
- 32:52make this also your backup plan. Thank you so much for being on
- 32:54the show. Is there anything else that
- 32:56you'd like to leave the audience with today about go to market
- 32:59data or anything like that? I want I want to add one thing
- 33:03that is currently happening a lot and I would just like to
- 33:09give advice to the listeners on like what not to do.
- 33:14Yes. Tell us is I write a lot about
- 33:18like what not to do and one of them is I want to definitely
- 33:21mention this personalization at scale.
- 33:24That's a good one, so. What does personalization at
- 33:26scale mean, right? You put AI on the data you have
- 33:30in your CRM or whatever you have get this data from and then tell
- 33:35AI to personalize the e-mail, right?
- 33:37So what happens is what is personalization?
- 33:41Let's start with that. Like why do we even personalize
- 33:44with personalization? We want to show the the prospect
- 33:49that we research their business, we understand their pain points
- 33:53and we know their priorities. We learn their business model
- 33:58their challenge is everything related to their business or
- 34:02specifically to their role. Now it's scale, AI,
- 34:05personalization. What does it do Here is an
- 34:07example that I got an e-mail hate about great posts about
- 34:14clean data. We think that our our medical
- 34:19devices will be a great fit for E core like medical devices will
- 34:24be great fit for E core. They personalized right.
- 34:27They saw my post of my post and they think that's enough like
- 34:31seriously medical devices so not to end up and I have so many of
- 34:36those gems like I like hundreds of this in my inbox and our
- 34:42inboxes are getting filled of the emails that just sound the
- 34:45same and it's oric. So it shows that there is a poor
- 34:49strategy that there is poor data and that's poor AI usage.
- 34:54So to avoid that, do not apply AI at scale for personalization.
- 35:01Personalization. Instead of blasting thousands of
- 35:04emails a day, send any that are really personalized because that
- 35:10will save your reputation. And for me, reputation is the
- 35:14key. I do care about my company's
- 35:16reputation and I know there are a lot others that also do.
- 35:20That's right. So that would be one really
- 35:22important advice I would like to give your listeners.
- 35:26OK, awesome. Thank you so much, Tamar.
- 35:29Really appreciate you having having been on the show.
- 35:32And to all our listeners, if you like this episode, drop us a
- 35:35comment, hit the subscribe button.
- 35:37Otherwise we will see you next time.
- 35:39Thank you so much. Thank you so much for having me
- 35:42or I appreciate that and thanks to your listeners.
- 35:45I hope they learn something new. Absolutely.
- 35:50Today's episode is sponsored by Computer by Devrev.
- 35:53Computer is enterprise AI built on one idea that shared memory
- 35:57makes AI work for go to market teams.
- 36:00It connects your systems, conversations and deal history
- 36:03into one living, breathing digital memory so your teams get
- 36:07accurate, trusted answers and can take action with confidence.
- 36:10Because Computer curates the right context before it ever
- 36:13hits the model, you consume up to 60% fewer tokens with every
- 36:17call. Find out more and start your
- 36:19free trial at Rev dot AI.