Tracking Offline Advertising Effectivness

by Joe Ray 32 replies
If the advertising is not online, not programmatic advertising and not a direct response ad, there is no way to know which ad campaign is causing the a conversions.

I built a Machine Learning solution for this problem. I wrapped this ML model into a self-contained, downloadable software that works on any PC.

Now what? How do I market this?

I have something in-between a prototype and a MVP at this stage. It works. I have been providing Data Science services but never created or launched a product.

Any suggestions would be appreciated.
#offline marketing #advertising #effectivness #offline #tracking
Avatar of Unregistered
  • Profile picture of the author Oziboomer
    Originally Posted by Joe Ray View Post

    Any suggestions would be appreciated.
    Hi Joe,

    I'd try to talk to Ewen Mack on this.

    You can look him up here in the forum and PM him to alert him to your post or he could be good to pay to consult on this as he has the experience in marketing to help from several angles when it comes to your situation.

    Best regards,

    Ozi
    {{ DiscussionBoard.errors[11112883].message }}
    • Profile picture of the author Joe Ray
      Ho Ozi,

      I had a conversation with Ewen a couple of years ago about something else but it didn't work out.

      Thanks for the suggestion anyway.

      Cheers,

      Joe
      Signature
      If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

      OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

      {{ DiscussionBoard.errors[11113248].message }}
  • Profile picture of the author Ron Lafuddy
    Originally Posted by Joe Ray View Post

    If the advertising is not online, not programmatic advertising and not a direct response ad, there is no way to know which ad campaign is causing the a conversions.

    I built a Machine Learning solution for this problem. I wrapped this ML model into a self-contained, downloadable software that works on any PC.

    Now what? How do I market this?

    I have something in-between a prototype and a MVP at this stage. It works. I have been providing Data Science services but never created or launched a product.

    Any suggestions would be appreciated.
    I read your headline: Tracking Offline Advertising Effectivness

    Doesn't seem like you are talking about offline marketing which is fairly easy to track.

    Am I missing something?

    Ron
    {{ DiscussionBoard.errors[11112973].message }}
    • Profile picture of the author Joe Ray
      Yes, maybe you're missing something.

      Let's say you're selling cars. You pay for some offline advertisement i.e. TV, Radio and Newspapers. Then, costumers come in to your dealership and buy cars from you.

      You have no way of knowing which ad (TV, Radio or Newspaper) actually caused the sale.

      This is what I mean by Tracking Offline Advertising Effectiveness.

      If you think it's "fairly easy" to track this, please explain what you mean.

      Oh, BTW I hope you don't mean surveys, because the sales people will not ask everyone who wonders into the dealership the "where did you hear about us" question and keep accurate data of the answers.

      I hope you get what I mean, if you walk into a Volkswagen dealership, they're not going to ask you where you heard about Volkswagen.

      I am very curious to find out what you meant by "fairly easy".
      Signature
      If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

      OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

      {{ DiscussionBoard.errors[11113259].message }}
      • Profile picture of the author Ron Lafuddy
        Originally Posted by Joe Ray View Post

        Yes, maybe you're missing something.

        Let's say you're selling cars. You pay for some offline advertisement i.e. TV, Radio and Newspapers. Then, costumers come in to your dealership and buy cars from you.

        You have no way of knowing which ad (TV, Radio or Newspaper) actually caused the sale.

        This is what I mean by Tracking Offline Advertising Effectiveness.

        If you think it's "fairly easy" to track this, please explain what you mean.

        Oh, BTW I hope you don't mean surveys, because the sales people will not ask everyone who wonders into the dealership the "where did you hear about us" question and keep accurate data of the answers.

        I hope you get what I mean, if you walk into a Volkswagen dealership, they're not going to ask you where you heard about Volkswagen.

        I am very curious to find out what you meant by "fairly easy".
        Yeah, fairly easy.

        Dealers run car ads on the back page of the newspaper, in the auto trader, or direct mail, etc.

        Cars are listed, with a particular price - and associated stock number for each car.

        When the customer comes in inquiring about the deal they saw on a particular model they refer to the stock number listed in the ad.

        Nothing new. Auto dealers have been doing this for years.

        Your welcome.

        Ron
        {{ DiscussionBoard.errors[11113294].message }}
        • Profile picture of the author Joe Ray
          What about TV and Radio? Or, just simple printed ad in a newspaper or magazine (which is what I meant by Newspaper ad)?
          Signature
          If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

          OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

          {{ DiscussionBoard.errors[11113346].message }}
  • Profile picture of the author SARubin
    Originally Posted by Joe Ray View Post

    If the advertising is not online, not programmatic advertising and not a direct response ad, there is no way to know which ad campaign is causing the a conversions.

    I built a Machine Learning solution for this problem. I wrapped this ML model into a self-contained, downloadable software that works on any PC.

    Now what? How do I market this?

    I have something in-between a prototype and a MVP at this stage. It works. I have been providing Data Science services but never created or launched a product.

    Any suggestions would be appreciated.
    Hi Joe,

    Pardon my ignorance, but I'm a bit confused about what you're software actually does.

    Perhaps I'm not reading your post correctly?

    You say, if the advertising is not online, and is not a direct response ad, then your software will allow me to track and measure which ad is directly responsible for the sale?

    But being able to track and measure which ad is "directly responsible" for a sale is a large part of the very definition of "direct response."

    And in direct response there will be a tracking mechanism built directly into the promotion. (tracking #, coupon, special dedicated phone line or special web page to visit, etc. etc.)

    I'm sure you already know that.

    So, if I put out a TV, Radio, Newspaper, and magazine ad, all at the same time, that all say the same thing, how does your software track that type of Institutional or brand awareness advertising, that has no tracking mechanism in play?

    (Without needing to spend an inordinate amount of time with data entry combined with guesswork every time a sale is made, or asking survey questions to every customer?)

    I would like to help you out here, but without knowing what your software actually does, it's a bit difficult to offer any suggestions.

    It sounds great, but most people who might be interested in what you're offering (including myself) are going to want to know a bit more about how it works, before spending money on it.

    Any chance you could give a little more explanation of what your program actually does (without giving away any of your trade secrets, of course)
    Signature
    Professional Sales Copy - For small business owners who are spending money on advertising, but are not getting enough sales (and would like to get more sales)
    {{ DiscussionBoard.errors[11113309].message }}
    • Profile picture of the author Joe Ray
      Hi SARubin,

      Thanks very much for your very educated and intelligent reply. I appreciate your interest in my product.

      Yes, you're right. It’s not possible to determine causation because we don’t have access to every possible external factor that might have influenced sales. Thus, our Machine Learning model focuses on association, rather than causation.

      What does that mean? We designed a highly complex algorithm to tackle this problem. Applying the algorithm, we built a Machine Learning model to look for hidden patterns in your ad spending and sales data.

      The output is a prediction which will tell you that, for example, a given amount of TV, Radio and Newspaper ad spending, an X amount of increase in, let's say, Radio ad spending will be associated with a Y amount of "unit" increase in Sales.

      More clearly: based on previous ad spending and unit sales data, we can predict which media is more likely to be associated with the sales or conversions.

      We tested the model and it works. The math involved in the model is not trivial. We still need to do some more testing and tweak the model if needed. But, it definitely works.

      Please let me know your thoughts.
      Signature
      If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

      OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

      {{ DiscussionBoard.errors[11113356].message }}
      • Profile picture of the author Claude Whitacre
        Originally Posted by Joe Ray View Post

        Hi SARubin,

        Thanks very much for your very educated and intelligent reply. I appreciate your interest in my product.

        Yes, you're right. It’s not possible to determine causation because we don’t have access to every possible external factor that might have influenced sales. Thus, our Machine Learning model focuses on association, rather than causation.

        What does that mean? We designed a highly complex algorithm to tackle this problem. Applying the algorithm, we built a Machine Learning model to look for hidden patterns in your ad spending and sales data.

        The output is a prediction which will tell you that, for example, a given amount of TV, Radio and Newspaper ad spending, an X amount of increase in, let's say, Radio ad spending will be associated with a Y amount of "unit" increase in Sales.

        More clearly: based on previous ad spending and unit sales data, we can predict which media is more likely to be associated with the sales or conversions.

        We tested the model and it works. The math involved in the model is not trivial. We still need to do some more testing and tweak the model if needed. But, it definitely works.

        Please let me know your thoughts.
        My experience is that most retailers and businesses advertise in multiple media at the same time. Is there a way to differentiate which media is creating the sales?

        Originally Posted by Joe Ray View Post

        I hope you get what I mean, if you walk into a Volkswagen dealership, they're not going to ask you where you heard about Volkswagen.

        I agree. But I think it's a good idea to ask what brought them into the store...in other words, was it an ad, a referral, or a sign they saw. The branding ads generally don't bring them in the door, it's the local ads that have model and price. That's my experience and the experience of my clients.


        In my retail store, when they buy, I ask what caused them to come in and shop in my store that day. I don't give them any prompts. I ask them because I often run multiple ads in multiple media at the same time.

        I only ask if they bought. And it has to be $300 or more.
        {{ DiscussionBoard.errors[11113947].message }}
  • Profile picture of the author SARubin
    Hi Joe,

    It sounds interesting (even fascinating)

    The first thing we need to do is to figure out who your target market is.

    Without spending a ton of time on researching the market, right off the top of my head I have a couple thoughts that might help point you in a target direction.


    First thought - Large (institutional type) advertising agencies are probably not a good market for you. Here's why...

    Most of them don't want to be held accountable for actual sales numbers. They prefer to sell the idea of "brand recognition / awareness" to their clients.

    They create clever ads that might win awards, and get people talking (or not) and if sales happen to go up, then they take credit for being brilliant.

    If sales don't go up, then they tell their clients to just keep spending ad dollars until they eventually become a recognized name (or until they run out of money)

    If most institutional type ad agencies where held accountable for actually producing measurable results, many of them would be out of business very quickly. Most of them know it, so they probably don't want your software holding them accountable.


    Second thought - Direct response advertisers could be a possible secondary market.

    Speaking from experience, I'm always interested in ways to track and measure results. And I could be interested in your software's ability to expand my services into new directions (assuming it actually works).

    On the flip side of that, I'm already fully committed to the concept of putting the tracking mechanism directly into the promotions themselves. So, you'd need to convince us that we even need what you're offering. (hence, the "possible" secondary market)


    Third thought - That brings us to who could be your primary market. Small to medium size businesses that are already spending money on institutional type advertising.

    These could be the people who are interested in "finally" being able to determine if their advertising dollars are being spent wisely, or if it could be spent more efficiently.

    Right now, many of them only have the ability to do what "normal" advertising tells them to do.

    Basically, spray and pray. Meaning, put out advertisements that seem good, and then see if their market share / sales go up.

    If the numbers go up then the ads were successful. If the numbers don't go up, then throw some more poo against the wall and see if something sticks. (or just give up because advertising doesn't really work anyway, right?)

    So I think this third category could be you primary target. But here's where the research begins. You'd need to start figuring out which companies are in the position to need your program.

    It would likely be companies that ...

    1 - spend money on advertising (obviously)

    2 - keep constantly changing their ads (could indicate they're not getting great results so they're still trying to figure out what works)

    3 - can afford what you're selling (here's where price comes in. For $30 you could sell it to small mom-n-pop businesses across the board. For $3000 you'll probably need to focus on slightly bigger companies, and offer some type of premium support, or full training of how to use your program efficiently, as part of the package)


    Big businesses could also be a viable market (especially the ones that have in-house marketing departments), but many big companies already use big advertising agencies who will actively try to sabotage you, because of what I already mentioned in the first thought above.

    So before going after the big boys, I'd recommend getting a few success stories under your belt first. That way you can more effectively counter any attacks that get thrown at you.


    Of course, this is all just off the top of my head. I haven't given it a ton of thought yet (and I only got about 3 hours of sleep last night... so that's a thing)

    Anyway, hopefully this gives you a couple things to think about. And maybe get your creative marketing juices flowing

    All the best,
    SAR
    Signature
    Professional Sales Copy - For small business owners who are spending money on advertising, but are not getting enough sales (and would like to get more sales)
    {{ DiscussionBoard.errors[11113888].message }}
    • Profile picture of the author Joe Ray
      Thanks so much for your valuable feedback. This is helping me a lot! I really appreciate it.

      Please see my notes below:
      First thought - Large (institutional type) advertising agencies are probably not a good market for you. Here's why...

      Most of them don't want to be held accountable for actual sales numbers. They prefer to sell the idea of "brand recognition / awareness" to their clients.

      They create clever ads that might win awards, and get people talking (or not) and if sales happen to go up, then they take credit for being brilliant.

      If sales don't go up, then they tell their clients to just keep spending ad dollars until they eventually become a recognized name (or until they run out of money)

      If most institutional type ad agencies where held accountable for actually producing measurable results, many of them would be out of business very quickly. Most of them know it, so they probably don't want your software holding them accountable.
      Your comment about the advertising agencies is an eye opener for me.

      You’re probably absolutely right about this. It makes perfect sense that they take the money and spend it offline without measurable results.

      I sent emails to a few ad agencies (less than 10) and offered them a free review copy of the MVP to test it and see if it covers any of their "pain points". No wonder none of them answered.

      But, this is bad news for me, because big ad agencies are/were my primary target. Now, I have to rethinking this approach.

      Second thought - Direct response advertisers could be a possible secondary market.

      Speaking from experience, I'm always interested in ways to track and measure results. And I could be interested in your software's ability to expand my services into new directions (assuming it actually works).

      On the flip side of that, I'm already fully committed to the concept of putting the tracking mechanism directly into the promotions themselves. So, you'd need to convince us that we even need what you're offering. (hence, the "possible" secondary market)
      If your method of (“putting the tracking mechanism directly into the promotions themselves”) works, you might not need my product.

      If you can pin down a problem by simply “counting”, you should probably not use Machine Learning (ML) algorithms.

      ML will still work, but if you can get an exact measure from your method, ML may not be optimal.

      My solution will make a huge difference for those who have no way of measuring the results for their offline ad campaigns i.e. TV, Radio, Magazines, Billboards.

      Third thought - That brings us to who could be your primary market. Small to medium size businesses that are already spending money on institutional type advertising.

      These could be the people who are interested in "finally" being able to determine if their advertising dollars are being spent wisely, or if it could be spent more efficiently.

      Right now, many of them only have the ability to do what "normal" advertising tells them to do.

      Basically, spray and pray. Meaning, put out advertisements that seem good, and then see if their market share / sales go up.

      If the numbers go up then the ads were successful. If the numbers don't go up, then throw some more poo against the wall and see if something sticks. (or just give up because advertising doesn't really work anyway, right?)

      So I think this third category could be you primary target. But here's where the research begins. You'd need to start figuring out which companies are in the position to need your program.

      It would likely be companies that ...

      1 - spend money on advertising (obviously)

      2 - keep constantly changing their ads (could indicate they're not getting great results so they're still trying to figure out what works)

      3 - can afford what you're selling (here's where price comes in. For $30 you could sell it to small mom-n-pop businesses across the board. For $3000 you'll probably need to focus on slightly bigger companies, and offer some type of premium support, or full training of how to use your program efficiently, as part of the package)


      Big businesses could also be a viable market (especially the ones that have in-house marketing departments), but many big companies already use big advertising agencies who will actively try to sabotage you, because of what I already mentioned in the first thought above.

      So before going after the big boys, I'd recommend getting a few success stories under your belt first. That way you can more effectively counter any attacks that get thrown at you.
      Yes, this seems to be my target market and this is where the marketing problem begins for me.

      This is a hard problem.

      I was very excited about targeting the big advertising agencies; because I am sure I can find a list of agency contact info and start pitching them the product.

      This seemed like the easiest way get to the potential buyers. The ad agencies have a lot of money and they can easily understand the product. Thus, if I can solve a problem for them I can easily make some sales.

      I wanted to price the product at $9,700 and sell it to about 100 agencies worldwide. This approach seemed like a reasonable one until I read your post. Now, I am a little bit discouraged about pitching to the ad agencies.

      But, I have a couple of concerns about approaching small to medium size businesses as well. 1.) They might not understand what I am offering exactly. 2.) They don’t want to spend the money.

      That being said, some good reviews about my product from a couple big brands would solve these problems very quickly.

      Thanks again for your immensely valuable input. I hope we can continue this conversation. Your feedback is encouraging and very helpful.

      I appreciate it very much.
      Signature
      If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

      OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

      {{ DiscussionBoard.errors[11115907].message }}
      • Profile picture of the author SARubin
        Originally Posted by Joe Ray View Post


        Your comment about the advertising agencies is an eye opener for me.

        You’re probably absolutely right about this. It makes perfect sense that they take the money and spend it offline without measurable results.

        I sent emails to a few ad agencies (less than 10) and offered them a free review copy of the MVP to test it and see if it covers any of their "pain points". No wonder none of them answered.

        But, this is bad news for me, because big ad agencies are/were my primary target. Now, I have to rethinking this approach.



        I was very excited about targeting the big advertising agencies; because I am sure I can find a list of agency contact info and start pitching them the product.

        This seemed like the easiest way get to the potential buyers. The ad agencies have a lot of money and they can easily understand the product. Thus, if I can solve a problem for them I can easily make some sales.

        I wanted to price the product at $9,700 and sell it to about 100 agencies worldwide. This approach seemed like a reasonable one until I read your post. Now, I am a little bit discouraged about pitching to the ad agencies.
        Hi Joe,

        I'd say you're right, most small business owners probably won't want to spend that kind of money. Although, companies that generate at least $10,000,000 or more, in gross revenue (still considered small business) might be interested?


        As far as the big ad agencies go, don't get discouraged. While most of them don't want to be held accountable for measurable results, that doesn't mean we can't approach this from a different angle.

        I'm forever the optimist, and your task may not be impossible (just a bit of a challenge).

        Of course, I'd never recommend spending too much time going after a market that doesn't want your product, but if you've only pitched to 10 big agencies so far, then you haven't even tested any "real" numbers yet.

        Now, I haven't seen the emails you've sent them, but if you're telling them about how to solve a problem they might not want solved, perhaps we just need a different approach?

        One that's less about pain points, and more about dominating their market could be worth testing. If they can show measurable results for their clients, or at least show scientific proof that their advertising methods can produce better results, it could set them high above other ad agencies (in a big way).

        Explain how this could be a game changer for them, and could solidly plant them ahead of their competition in the advertising industry. Throw in some talk about AI and machine learning (which is all the buzz these days) and you might be able to peak their interest.

        So, don't give up on the big agencies just yet. Instead start thinking about how your software can help them dominate "their" market. And how they might want to get in on this before their competition does.

        Something to think about.

        All the best,
        SAR


        P.S. One last thought... We need to test more than just 10 cold emails before getting discouraged.

        10 emails is hardly enough to measure results (unless you've spent a lot of time studying, or building a relationship with, the ten people you sent the email to and you know exactly what they want, or need)

        Also, do you know if your emails even got opened? Maybe it went to a spam folder, or your subject line caused them to hit the delete button without even opening it?
        Signature
        Professional Sales Copy - For small business owners who are spending money on advertising, but are not getting enough sales (and would like to get more sales)
        {{ DiscussionBoard.errors[11115982].message }}
      • Profile picture of the author savidge4
        Originally Posted by Joe Ray View Post

        I wanted to price the product at $9,700 and sell it to about 100 agencies worldwide. This approach seemed like a reasonable one until I read your post. Now, I am a little bit discouraged about pitching to the ad agencies.
        Its as easy/hard to sell 100 as it is 10,000. The thing is there are only so many ad agencies in the world, and there are infinite small business'. So you could target agencies and struggle to sell 100, or you could sell 10,000 at $97 to small business owners, that struggle with optimizing their advertising budget.

        Looking at this from a commerce perspective if you were to sell to the 100 agencies you would stop at 100 no? if you target small business 10,000 maybe the target, but automating the system.. over time, you could sell 10K, 20K, 50K. And it doesn't have to be $97 a pop.. it could be $197 or $497. Pricing is something you can test, find the point of diminishing returns.

        You are looking for a finite, when infinite is possible.

        You are solving a universal pain point for each and every small business owner.
        Signature
        Tools, Content, and Strategies to Develop Your Online Commerce Business - Coming Soon
        {{ DiscussionBoard.errors[11116543].message }}
        • Profile picture of the author Joe Ray
          Yes, I am still open to all options, as far as target market and pricing.

          Your suggestion to lower the price and target a wider pool of potential buyers makes sense.

          I also considered this option and I am still undecided on which way to go.

          Here is my logic for targeting fewer buyers with a higher priced product:

          Offline adds, especially TV ads or billboards, are not cheap. In some cases, they are very expensive. Thus, I think, the target market for this product is most likely companies with a lot of money.

          And, to narrow it down even more, I probably should look for companies who are consistently spending a lot on their offline campaigns. Not just occasionally.

          Why? Because, for these companies, even a 70% sales prediction accuracy, can produce millions of dollars of extra revenue.

          In these cases, even the $10K unit price for my product is actually very low.

          That being said, I would like to explore your suggestion as well.
          Its as easy/hard to sell 100 as it is 10,000. The thing is there are only so many ad agencies in the world, and there are infinite small business'. So you could target agencies and struggle to sell 100, or you could sell 10,000 at $97 to small business owners, that struggle with optimizing their advertising budget.
          My issue is that I've only done high priced, B2B product sales before. That means, I can see how I could reach 100 buyers, but I can't wrap my head around how I can possibly reach 10,000 buyers.

          I know it's doable, I just haven't tried this before.
          Signature
          If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

          OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

          {{ DiscussionBoard.errors[11120118].message }}
          • Profile picture of the author savidge4
            Originally Posted by Joe Ray View Post

            Offline adds, especially TV ads or billboards, are not cheap. In some cases, they are very expensive. Thus, I think, the target market for this product is most likely companies with a lot of money.
            Those 2 in particular with the exception of newspaper... are probably the cheapest... I find that radio is by far the most expensive. I know this from personal ( business ) experience. I am not going to say my business does not make money, but I wonder if it qualifies for a business you think makes "a lot of money"

            How many retired rich restaurant owners do you know? To be honest I know 2.. and I know a lot of restaurant owners... they are far from rich, but you see them regularly on TV and Billboards no?

            One of my offline business' is Satellite TV and Internet. I actually get CO-OP funds to pay for offline ads ( IE Newspaper, Billboards, Radio, and TV ads. ), so the expense isn't so much a concern.. but I do want to use the money wisely. Companies like Hughesnet, or Perfect 10 or Exede or Dish Network would probably buy your App. so maybe 6 copies... and there are 1000's if not 10's of thousands of independent distributers / retailors. The point being, Be it agencies or CO-OP programs or any other thing... sure you can sell to the top of the pyramid, but there are endless amounts of potential customers below them.

            Something you might consider... On Exede's website ( exede.com ) under "Quick Links" ( mid page left side ) you could click on "Find a local dealer" and by typing in a zipcode, you would get a list of all the dealers. You could with a little work, develop a hell of a "List" and more importantly you would be developing a list of buyers, that are not focused so much on the expense ( they ALL have access to CO-OP dollars ) and are more concerned with results.
            Signature
            Tools, Content, and Strategies to Develop Your Online Commerce Business - Coming Soon
            {{ DiscussionBoard.errors[11120730].message }}
            • Profile picture of the author Joe Ray
              I checked the ( exede.com ) site. This is a very interesting idea. Definitely good suggestions.

              And, sounds like you could also use my app.

              If you're interested and you think my app might solve a problem for you, I am happy to offer you a free review copy for feedback and some marketing help i.e. a good review etc.

              BTW: You've already helped me a lot with your suggestions and encouraging posts.
              Signature
              If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

              OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

              {{ DiscussionBoard.errors[11125289].message }}
  • Profile picture of the author yukon
    Originally Posted by Joe Ray View Post

    If the advertising is not online, not programmatic advertising and not a direct response ad, there is no way to know which ad campaign is causing the a conversions.

    Joe you're wrong and making a simple task complicated.

    Coupon codes.

    You can track anyone with a unique code.
    Signature
    Hi
    {{ DiscussionBoard.errors[11113961].message }}
  • Profile picture of the author savidge4
    Originally Posted by Joe Ray View Post

    If the advertising is not online, not programmatic advertising and not a direct response ad, there is no way to know which ad campaign is causing the a conversions.

    I built a Machine Learning solution for this problem. I wrapped this ML model into a self-contained, downloadable software that works on any PC.

    Now what? How do I market this?

    I have something in-between a prototype and a MVP at this stage. It works. I have been providing Data Science services but never created or launched a product.

    Any suggestions would be appreciated.
    Let me preface this a bit.. I went to your site and looked at your "New Products" write up on this. I have an engineering degree I get it. This is potentially game changing stuff here. BUT... only because I have an engineering / math background do I get it.

    What you REALLY need to do... find 5 "test" subjects. Small business owners that you can sit down and input the data for them, and let your program spit outs its data. THEN listen to them explain to you what that translates to in business talk. Right now there is simply a language barrier between what you are saying, and what a average business person is able to hear.

    How I see this.... "Big Data brought Down to size" In the advertising world it is understood that online advertising has ability to track where and when conversions take place. This many people went to your site. This many people looked at your facebook page. That many people bought an item from your site. Easy to understand numbers.

    Your offline marketing IE TV ads.. Radio Ads... Newspaper Ads.. not so easy. Instead of tracking actual conversions you get numbers like 86% of this demographic are reached daily. Your message is reached across 12 Counties. There are X amount of readers that will see your ad. Nothing says "we will provide X% of conversions from these numbers. And these numbers sound really good right? but how do you know you are spending your ad dollars wisely?

    With our "Predictive Analytics" all of that changes. Before you spend another dollar with offline advertising... KNOW what your ROI could be. Simply start up the self contained app on any PC insert the numbers, and let the Market specific AI goto work, and return REAL potential results from Newspaper ads, from Radio Adds, from TV ads.

    You no longer are hoping that it works.. being told.. it takes months for results. You will have the needed data right in front of you ( before you actually make the spend ) to not only make a decision of where to spend your ad dollars, but where those dollars are going to work best for you.

    Best of luck with this one... I personally see potential here.. good stuff!
    Signature
    Tools, Content, and Strategies to Develop Your Online Commerce Business - Coming Soon
    {{ DiscussionBoard.errors[11115163].message }}
    • Profile picture of the author Joe Ray
      Let me preface this a bit.. I went to your site and looked at your "New Products" write up on this. I have an engineering degree I get it. This is potentially game changing stuff here. BUT... only because I have an engineering / math background do I get it.
      Thanks very much for your comment. I appreciate your positive feedback. This is really helpful and encouraging.

      What you REALLY need to do... find 5 "test" subjects. Small business owners that you can sit down and input the data for them, and let your program spit outs its data. THEN listen to them explain to you what that translates to in business talk. Right now there is simply a language barrier between what you are saying, and what a average business person is able to hear.
      Yes, I am still looking for “test” subjects. I found one a couple of weeks ago. I plugged in their data to my software and they were shocked when they saw the results.

      It would take too long to explain here how we did the test exactly. But, the company was really impressed with the result.

      They already had a hypothesis about which spending was producing an increase in sales and which spending was not working at all. My predictions definitely supported their hypotheses.

      This experience provided a positive reinforcement for me, because before this test I only tested the model on one data set. It was certainly encouraging to see that it produces the expected results on completely new data. The company will write a positive review, recommending my software. This should boost my marketing effectiveness.

      But, you’re right! There is still a language barrier when I try to market this solution. I’ve already spent about a month trying to produce good marketing material but I am not happy with what I have so far.

      BTW I rather write hundreds of lines of code than 10 lines of sales copy. I am not a sales copy writer and I don’t have an inspiration to ever become one. Yet, I still have to sell my product somehow.

      How I see this.... "Big Data brought Down to size" In the advertising world it is understood that online advertising has ability to track where and when conversions take place. This many people went to your site. This many people looked at your facebook page. That many people bought an item from your site. Easy to understand numbers.

      Your offline marketing IE TV ads.. Radio Ads... Newspaper Ads.. not so easy. Instead of tracking actual conversions you get numbers like 86% of this demographic are reached daily. Your message is reached across 12 Counties. There are X amount of readers that will see your ad. Nothing says "we will provide X% of conversions from these numbers. And these numbers sound really good right? but how do you know you are spending your ad dollars wisely?

      With our "Predictive Analytics" all of that changes. Before you spend another dollar with offline advertising... KNOW what your ROI could be. Simply start up the self contained app on any PC insert the numbers, and let the Market specific AI goto work, and return REAL potential results from Newspaper ads, from Radio Adds, from TV ads.

      You no longer are hoping that it works.. being told.. it takes months for results. You will have the needed data right in front of you ( before you actually make the spend ) to not only make a decision of where to spend your ad dollars, but where those dollars are going to work best for you.
      This is tremendously helpful! Thanks so much for taking the time to write this.

      Best of luck with this one... I personally see potential here.. good stuff!
      As I mentioned, your reply is giving me a lot of positive energies. That doesn’t happen often on this forum. Thanks again!
      Signature
      If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

      OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

      {{ DiscussionBoard.errors[11117256].message }}
      • Profile picture of the author Oziboomer
        Originally Posted by Joe Ray View Post

        But, you’re right! There is still a language barrier when I try to market this solution. I’ve already spent about a month trying to produce good marketing material but I am not happy with what I have so far.

        BTW I rather write hundreds of lines of code than 10 lines of sales copy. I am not a sales copy writer and I don’t have an inspiration to ever become one. Yet, I still have to sell my product somehow.
        Have you thought about doing JV with someone who can market the solution for you?

        You could reach out to someone like Ryan Deiss at DigitalMarketer. Over the years they have partnered with people to bring products to market.

        They are highly active and have numerous contacts and the ability to run some serious campaigns.

        They understand the big data picture and would have some serious data themselves to run through the system.

        You'd be partnering with some of the best copywriters and marketers out there.

        They did used to have an application form somewhere for people to submit requests but you could just reach out and ask.

        Even if they weren't interested the connection may be helpful and they may refer you to someone else. Ryan had always mentioned hitting him up on twitter was the quickest way to get his attention but a call to his office and a fedex-ed pitch might open the door.

        If you do contact him tell him the "guy with the boomerang" said it would be OK. He'll know who you mean.

        They certainly have the knowledge and capacity to take things large if that is what you desire.

        Best regards,

        Ozi
        {{ DiscussionBoard.errors[11118535].message }}
  • Profile picture of the author umc
    I've been reading this with interest because I like to learn. Obviously you are going to have to be able to explain what your product does that can't already be done. It seems to have been established that pretty much anything can be tracked already. In the end, won't you have to have some of that hard data for your "machine" to work? You'll have to know the efficacy of the advertising being done by this particular company in order to predict anything. If they have those numbers, clearly they're already tracking their numbers, so why do they need you? If they don't have those numbers, then they wouldn't have anything for your machine to work with or base anything on.

    What am I missing? I'm sincerely trying to grasp the concept. Maybe I don't have a grasp of higher level agencies and what they do. I would think that they have ways to track already though, otherwise they wouldn't be higher level agencies.
    {{ DiscussionBoard.errors[11116123].message }}
    • Profile picture of the author SARubin
      Originally Posted by umc View Post

      I've been reading this with interest because I like to learn. Obviously you are going to have to be able to explain what your product does that can't already be done. It seems to have been established that pretty much anything can be tracked already. In the end, won't you have to have some of that hard data for your "machine" to work? You'll have to know the efficacy of the advertising being done by this particular company in order to predict anything. If they have those numbers, clearly they're already tracking their numbers, so why do they need you? If they don't have those numbers, then they wouldn't have anything for your machine to work with or base anything on.

      What am I missing? I'm sincerely trying to grasp the concept. Maybe I don't have a grasp of higher level agencies and what they do. I would think that they have ways to track already though, otherwise they wouldn't be higher level agencies.

      .
      Hi UMC,

      Yes, for the "direct response" crowd, anything that can be tracked, is tracked.

      And in my perfect world, all advertising would be direct response.

      But in reality, a lot of it is not. Just go open any national magazine and look at the ads. Most of them are flashy "brand awareness" ads with pretty pictures and catchy taglines, but no direct call for action, or tracking mechanism.

      Also look at T.V. ads (although thanks to Netflix I rarely watch commercial T.V. anymore ) , billboards, most radio ads, etc. Clearly there's a lot of advertising that's not being tracked.

      What a lot of them do is try to saturate the market by putting out a lot of ads. Then measure increased market share, without ever knowing if it's the magazine, radio, billboard, T.V. (or a combination of two or more) that's causing the increased market reaction.

      So, I do see a lot of potential for this type of program.

      Personally, I'm not certain how this software actually works. But savidge4 says the math / engineering looks viable (and he knows a whole lot more about that type of math then I ever will) So that, combined with what I see as potential for the concept, makes me think this software is very interesting. (possibly game changing)

      Here's where the job of copywriting and marketing comes in. Someone will need to bridge the gap between the math / engineering of it all (the stuff that you, me, and most mortals don't quite understand) and the utilization benefits to the target market.
      Signature
      Professional Sales Copy - For small business owners who are spending money on advertising, but are not getting enough sales (and would like to get more sales)
      {{ DiscussionBoard.errors[11116172].message }}
  • Profile picture of the author umc
    I can wrap my head around there being large companies that put ads out without really tracking them, as often when any organization gets large it becomes inefficient, and like you said many of those ads are nothing but branding with no call to action. Heck, a lot of small companies are too lazy to track or just never though of doing so as well.

    I guess I can't see how software could be accurate in predicting the efficacy of marketing though without input from those specific campaigns to forecast what would be a likely outcome. You can't just use industry standards because that particular company may be extremely proficient or deficient in their messaging. In other words, everyone is different, so I'm not sure how without specific inputs from those companies you can garner insight into their marketing futures, so to speak.

    I get that people smarter than I can come up with viable things that I don't get, but if he wants money from these businesses he's going to have to make it understandable.

    I believe that earlier you said that you're a natural optimist, and I probably see this different because I'm a natural skeptic. So I can't just feel like something might be a game changer, I have to understand it. Savidge is a bright dude with his hand in every conceivable business, lol, and kind of my idol on here because in a different world I would have been that guy with his hand in everything, so I trust his input. I still want to grasp how this can work though, lol.
    {{ DiscussionBoard.errors[11116316].message }}
    • Profile picture of the author SARubin
      Originally Posted by umc View Post

      I guess I can't see how software could be accurate in predicting the efficacy of marketing though without input from those specific campaigns to forecast what would be a likely outcome.


      I believe that earlier you said that you're a natural optimist


      I still want to grasp how this can work though, lol.
      Yeah, I can't quite wrap my head around how it works either. But I do know that I'm interested in the possibilities of it.

      I think some of my current optimism comes from the books I've been studying lately about how to think like a visionary.

      The latest example talks about a guy who (some 40 years ago) envisioned a paperless office that no longer needed to use paper for keeping records (receipts, invoices, inventory records, etc.)
      Most people laughed and called him a dreamer, and told him to go back to college and train for a "real" job.

      Where would we be right now if Bill Gates had listened to all those naysayers?


      So, does this marketing software actually work? I don't know... yet.

      If it does work, then "how" does it work? I don't know... yet.

      Finally, can we explain it simply enough (and market it) to make companies want to buy it? I don't know... yet.

      If it can't be done, then I've lost nothing but a little misplaced enthusiasm. But if it can be done, then I'm fascinated with the potential of it.
      Signature
      Professional Sales Copy - For small business owners who are spending money on advertising, but are not getting enough sales (and would like to get more sales)
      {{ DiscussionBoard.errors[11116360].message }}
      • Profile picture of the author savidge4
        Originally Posted by SARubin View Post

        If it does work, then "how" does it work?
        If you goto the OP's site, hit "New Products" he actually lists a number of algo's that are used. Spend a few minutes on Wikipedia and read about each of them. I would say start with Random Forest.. then follow that with Linear Regression.

        I am only speculating - that they are calculating regressions, and filtering regression groups through filters, to achieve calculated finite resolutions - IE X spent on paper ads will result in Y% return.

        Think of it as calculated advertising risk analysis. As Marketers we want to think of ad success as headlines and copy and blah blah blah.. but the reality is even the not so right message in the right place at the right time will provide results.
        Signature
        Tools, Content, and Strategies to Develop Your Online Commerce Business - Coming Soon
        {{ DiscussionBoard.errors[11116554].message }}
  • Profile picture of the author Doug9
    The businesses you pitch this to are going to be skeptical. Like me. They are going to think to themselves, "If I buy a billboard, how in hell can he tell me the effectiveness of it and how much profit I made from it?"

    If you can PROVE it to them, then it's worth the $3000 subscription price. If its one of these things that are sometimes right and sometimes wrong, then you have a long slog ahead.
    {{ DiscussionBoard.errors[11117892].message }}
  • Profile picture of the author ewenmack
    Hi Joe,

    I'm seeing you are packaging a product/service that you are trying
    to sell without verifying there is a demand.

    I'll give you an example as to what is in demand
    on the surface which could be explored further
    to see if there is actual buyer intent.

    Last year I spoke with a lady who answers to the ceo of Vodafone NZ.

    In the discussion was mobile phone customer retention.

    It was hurting them.

    They were doing reviews over different departments
    and funding internal changes

    I suggested analyzing the data for signals for drop offs.

    She was interested and felt they didn't have the in-house knowledge to mine that data.

    See the difference here?

    A real conversation in a big company with a real problem
    voiced by her.

    I knew retention of customers was a hot topic for Vodafone NZ
    because it was in the media how many customers they had lost.

    This discussion hasn't been about a specific problem voiced by a potential client.

    Can you see the difference?

    BTW, I'll phone my contact at Vofafone NZ to see where there
    internal changes are at and see if there is an opportunity for
    using data mining to help them with customer retention.

    Best,
    Ewen
    {{ DiscussionBoard.errors[11125405].message }}
    • Profile picture of the author Joe Ray
      Hi Ewen,

      Preventing churn is usually apart of a CRM system. It's hard to imagine that Vodafone doesn't have a CRM that addresses the costumer churn problem.

      One of my projects is to build a CRM with an integrated Artificial Intelligence (AI) System. One of the most important functions in this CRM is the churn prevention. We use Deep Neural Networks to solve this problem.

      A while ago I posted some details on our CRM project on the WF blog here:
      http://www.warriorforum.com/blogs/jo...-problems.html

      BTW, this project is not founded yet. That's a big problem. It's expensive to build this kind of systems.

      Also, there are many companies building AI integrated CRM now. Big companies like Salesforce or Microsoft and many smaller companies.

      Again, I don't see why Vodafone wouldn't have one installed.
      Signature
      If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

      OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

      {{ DiscussionBoard.errors[11125435].message }}
  • Profile picture of the author ewenmack
    If they do, they don't know how to read it, [her words]
    then to make decisions to stop the churn.

    They lost around 60,000 customers in one year
    for a total population a bit over 4 mill.

    Best,
    Ewen
    {{ DiscussionBoard.errors[11125445].message }}
  • Profile picture of the author ewenmack
    [DELETED]
    {{ DiscussionBoard.errors[11125450].message }}
    • Profile picture of the author Joe Ray
      I would love to get a client like that with this type of problem! If you can get this project/client, we can definitely solve this problem for them.
      Signature
      If you need FREE Data Science - Artificial Intelligence consulting, send me a message here

      OR, Read my new blog post here on Warrior Forum about: Accelerating Complexity - Integrated AI Systems

      {{ DiscussionBoard.errors[11125464].message }}
  • Profile picture of the author Jefay
    Do you use some promotional items for marketing? How do you track your campaigns if yes?
    {{ DiscussionBoard.errors[11155258].message }}
  • Profile picture of the author Paul Faust
    Great thread...would love to learn more about your machine learning solution. My company has worked in this business for many years. We help companies track different media by providing unique True 800 tracking numbers. A different number can be used in each ad and with either our platform or another a company may currently use...you can track each call back to the media that ran.

    If anyone would like more information on tracking numbers, easy dial numbers or even memorable phone numbers you can visit http://www.ringboost.com and/or contact me on this thread.

    I look forward to seeing more posts on this topic.

    Paul Faust
    RingBoost.com
    {{ DiscussionBoard.errors[11156174].message }}
  • Profile picture of the author Jefay
    [DELETED]
    {{ DiscussionBoard.errors[11165069].message }}
Avatar of Unregistered

Trending Topics

  • 135 {{ upvoteCount | shortNum }}

    Tons of Bad Advice In This Thread

    Most people that come to the forum are looking for advice on how to build their business or to otherwise make more money. Thanks to all those willing to help ... [read more]

  • 2 {{ upvoteCount | shortNum }}

    Website Designing and Development

    Rajeev Gaur in Web Design

    Hi there, What is the difference between website development and website designing? thanks in advance

  • 7 {{ upvoteCount | shortNum }}

    How can we get good back links ?

    sellebrate in SEO

    Please tell me I need this one so much !

  • 1 {{ upvoteCount | shortNum }}

    Users for website to give appearance of activity

    HectorGonz in Growth Hacking

    Hello Everyone, I have a new website that I'll be launching soon. It will be a niche social media site. While I'm promoting the site through various methods, I'd like ... [read more]

  • 8 {{ upvoteCount | shortNum }}

    hi this is my first post

    babarmaqbool in Off Topic

    hi every one i am new here .