Marketing Blog

Google is Charging More for Tablets

May 11, 2013 – 7:13 AM

The other day I was looking over one the PPC accounts decided to look at campaigns by device. I typically do not look at things this way as the device for this particular account does not matter as much as other things. Now, with that said I still find it to be of value, but as a secondary metric.

When looking at the device type by traffic and spend I was somewhat surprised with the results. The avg. CPC by device type was not what I was expecting at all.

So it is broken down by 3 device types

Computers
Mobile Phones
Tablets

For this account a greater majority of traffic and spend is targeting computers. Mobile does not generate a lot for us, but it is in the thousands of clicks on a monthly basis.

Tablets are not a big traffic source, but for some reason I became fixated on the stats. I kind of look at tablets from a PPC point of view as an emerging device type that has stay power that I suspect will surpass mobile devices. Tablets a simple lightweight, easy to use, and more portable than a laptop or computer, but at the same time offers more functionality and usability than a phone which makes it a leading candidate for the eventual replacement of everyday computing.

Back to the PPC stuff…

What I found is that in every non-mobile specific campaign in this account, tablets avg. CPC was 5%-10% more than computer avg. CPC. Lets fast forward a few years and assume that the AdWords Algorithm stays the same, tablets will have a much larger market share than it does today and there will continue to be an erosion from the computer and laptop market share, vendors using AdWords might eventually end up paying 5%-10% more than they do now.

Perhaps it’s just me, but I do not think its fair to companies that invest in a service that has progressively become more expensive to use since the recession started to take emerging devices that will one day surpass the current device leader and charge more per click. The way that it is being done today is in such a deeply buried way that most of the people looking and working in Adwords accounts will not even know the difference. And if nobody is saying anything about this, that means that Google will move forward with a new CPC avg. that will eventually make them 5%-10% more money with a simple price hike that nobody can see.

I am all for making money and do not necessarily blame Google for doing this, but there needs to be a great equalizer in the search world that gives vendors and users of Google Adwords some kind of leverage to things like this because without any kind of protection, Google basically dictates exactly how much money and how successful companies can be and that is too much power and control for one entity.

By mgoldberg | Posted in Marketing | Tagged Analysis, Device Bidding, Google AdWords, PPC, Reporting, Segments, SEM | Comments (0)

Why Building a True Funnel In SalesForce is Problematic

May 9, 2013 – 8:07 PM

SalesForce is a CRM that can be your company’s best friend or worst enemy. I have a love hate relationship with SalesForce because from a sales data entry and management perspective, it’s flexible and easy for a sales team to use. In a nutshell SalesForce, if used right can be a powerful asset to any person that uses it.

But wait; there is that hate part of the relationship that I mentioned… If you are a report user of SalesForce you know all too well that connecting data between Campaigns, Leads, Contacts, Accounts, and Opportunities in a fluid way can be a real issue. You can pull in some data from some of the reports available, but you really can’t pull in ALL of the data from ANY of the SalesForce silos and that makes building a top down funnel really challenging for many people.

I have recently been tasked with building out a funnel for sales and marketing and I have multiple databases to gather data from which is fine, but pulling in SalesForce data in to a spreadsheet (that is my first choice) was basically me building a series of frankenreports.

What I mean by that is I needed to bring in the following information to the funnel:

Impression data from Google Analytics
Impression data from various PPC
Print magazine circulation size
Trade show attendance data
Other various 3^rd party data tied to marketing
Web visits from Google Analytics
Lead volume from SalesForce
Contact data converted from lead volume from SalesForce
Campaign data tied to some of the lead volume (partially mashed up to segment some lead volume)
Opportunity data
- Multiple stages in opportunities with record count, and revenue tied down the stages
- Closed won record count and revenue
- Closed lost revenue and record counts
- Cost associated with every indicated function and data point mentioned above

In other words, I am pulling 10 or more reports in to a central master report that houses everything in a spreadsheet. In this spreadsheet I have about 15 -20 metrics that are used by about 10 people.

Anyone who is reading this right now knows that what is mentioned above is a tall order and yes it can be done, but one really needs to know their data very well because when you mash up 10+ reports from 3 DB’s and not to mention validating data in a 4^th DB is a nightmare if there is poor connectivity or integration between databases.

Back to SFDC for a little bit… I decided that I will build 9 reports in SFDC that incorporate each step of the funnel from the lead record down. Sounds like a lot of work, but I am VERY comfortable with SFDC and after building 1 the rest took no time at all.

Because I cant present 9 reports to people and say it’s a funnel trust me, I built a dashboard in SalesForce and had 3 columns and each column had 3 funnel charts with the following data:

Column 1
- Funnel 1 = Lead to Contact record count with conversion rate built in to the funnel (this was EASY)
- Funnel 2 = Opportunities created from the above leads (we do not create opps and contacts at the same time which is why its done this way) by record count by stage with % of record in each stage
- Funnel 3 = closed lost opportunity by record count (again, this was simple to do)
- Column 2
  - Funnel 1 = Lead to Contact record count with conversion rate built in to the funnel (only there for visual representation down the column)
  - Funnel 2 = Opportunities created from the above leads by revenue by stage with % of record in each stage
  - Funnel 3 = closed lost opportunity by revenue
  - Column 3
    - Funnel 1 = Lead to Contact record count with conversion rate built in to the funnel (again, visual representation only)
    - Funnel 2 = Opportunity average sales price (I added a simple SFDC FX to the report)
    - Funnel 3 = closed lost opportunity average sales price (perfect for understanding deal size issue’s not closing…

After building this dashboard with a bunch of mini funnels, I made 80% or more of my data mining now fully automated. My preference is to find a plug-in or reach out to one of our expert in-house developers to sync the rest of the data, but I am OK if I have to throw it in a spreadsheet.

So I found a way to make a lot of what I am trying to build in a funnel easier, but with all due respect to the great people at SalesForce, most people that use SalesForce are not power users, not developers, not coders, and not as familiar with the inner workings of the way data for their company flows and as such they really need to find a way to have an OOB report option that pulls in data and all fields from every SFDC record type available and make it dashboard compatible and chart compatible. If they added that one feature, SalesForce would hands down be able to give companies a singular snapshot of a company’s health and be able to connect the dots in a significant and actionable way that anyone can understand no matter what level you are at.

I have seen a few appexchange funnels that are free and they do some of this, but honestly SalesForce needs this kind of functionality for reporting. More and more big data and SaaS is the essential framework of sales and marketing for business and fragmented data is NOT the way of the future.

By mgoldberg | Posted in Big Data, Marketing, Sales | Tagged Big Data, Frankenreports, Funnels, Marketing and Sales, SalesForce, SFDC | Comments (2)

Ecommerce – Cart Sessions vs Transactions

February 19, 2013 – 7:01 AM

Online Marketing covers such a wide swath when it comes to areas of focus, but not many can be as lucrative as your shopping cart. A well planed shopping experience can make huge increases in revenue and that should be what is on the mind of any marketer and company.

Now with that said, there are several metrics that one should look at when it comes to testing the effectiveness of your cart. I am going to dive in to a couple of these metrics.

Scenario:

You are running a split test (50/50 traffic) that exposes people to one of pages. The intention of this test is to clearly define a winning design that statistically sends more traffic from the entry page to the cart.

Note: Essentially this means that someone enters a landing page and clicks buy now.

What is the primary metric of success?

Simple enough, it’s the cart entrances (sessions). This is not a terrible metric; in fact I think it could be considered as a step one metric to a greater overall user experience. After all, at the end of the day you want to see more traffic to your cart and in theory the more traffic you send to the cart, the better your chances of getting a sale.

As you run the test the tool that you are using has determined that the new design is sending more traffic to a cart session by the tune of an observed difference of 70%. Your tool has declared a winner and you feel good about increasing cart sessions. It is game over, and its time to move on to the next great thing, right?

Looks might be deceiving

While having an impressive spike in cart sessions by using the new design, what about transactions? How did they do, lets not pop the cork on the champagne bottle until you check what I find to be the most important metric that any online marketer with a shopping cart needs to look at.

So in this scenario, all split tests can be tracked in Google Analytics, including the transactional data. What we discovered was that during the test that had this incredible observed difference of 70% on the newly designed landing page there was a 41% decline transactions. Transactions in this case = revenue.

Because the test was really intended for cart sessions and not sales nobody would even consider looking at that. After all, if you send more traffic to the cart, you should get more sales, right?

Now we know that is not always right, but the takeaway from this is that everyone should always understand that the most important metric any anyone who is involved in an online business is sales. Other metrics matter, but nothing as much as sales data that flows and feeds the company.

By mgoldberg | Posted in Marketing | Tagged A/B Testing, Analysis, Ecommerce, End User Sales, Google Analytics, Metrics, Shopping Cart | Comments (0)

Phone Sales Tip – Use Pricing as a Guide Not a Rule

February 17, 2013 – 10:09 AM

More often than not sales organizations have the propensity to handcuff their sales team with strict enforcement of pricing. While its always nice to get the max price for your company, I think that if you have a person that is on the phone and willing to buy now for X% off you should absolutely go for it.

There are so many valid reasons to bend on pricing, but the single most important reason that I can think of is plain and simple, revenue is revenue and if you could not get them at full price, are you really going to not do what’s best for the company (and your paycheck) because of a price book?

Now, if you have start cutting deals on all of your sales, that could be a problem, but I think almost every company, and even manager will gladly accept a sale that is not exactly according to a pricing guide. After all, management needs to project revenue and have annual goals, if they don’t hit the numbers for something like price adherence they could be doing themselves a disservice.

Several months ago I was asked to take on a team that I managed for years and it was managed by 4 other people. Each of these people had this mentality of what their team should do when it comes to pricing. Some were more forgiving than others, but they all shared one common process, do not deviate to far from whatever the price book says.

I will be the first to claim, I am not your standard sales leader and that’s mostly due to the fact that I came up from the trenches and know exactly how frustrating it is to be held to a number and not have the flexibility to do what ever is necessary to hit that number. As such, I have one simple rule with pricing that I enforce daily with my team: “Anything greater than zero is OK by me.” Now that’s a bit of an exaggerated statement, but the team fully understands this statement really means, do what you think is best for closing the sale and don’t worry about the price as much as closing business.

In my model, these sales are mostly one and done, but there is millions to be had from these people in reoccurring revenue one to two years down the road. So if you cut X% up front, you can make that up on the back end anyway, so the loss is really nonexistent.

The real loss would be to stay strict on a pricing guide that potentially prohibits overall growth and sales for the company, team, and your wallets.

By mgoldberg | Posted in Sales | Tagged Consumer Sales, End User Sales, Phone Sales, Sales Tip | Comments (2)

Keyword Spy and Other SEM BI Tools

February 1, 2013 – 6:34 AM

Tools like Keyword Spy, SEM Rush, and Spyfu have a lot of useful information when it comes to understanding the ways that some people show up for ads and organic rankings.

You can find out great things like who your competitors could be and how they segment and target their keywords. This alone is a value to any company that wants to know what is up with the competition.

However, there are some things about any of these services that are flat our incorrect. I was in a meeting recently and someone was talking about ad spend on one of these “spy” tools and this person even acknowledges that this is a baseline of data, not entirely accurate. Kind of a baseline more than anything else.

I can live with that, but if you have been dealing with the same competitors for a decade, you develop an instinctive gut check that can be trusted about who is spending more.

I have used keyword spy for a while now, but also use SEM Rush and Spyfu in the past. They all work under the same principal when it comes to PPC. In this particular instance I suspected that we spend more money on a segment of the business than others, but the actual “baseline” in these tools contradict my gut check.

I am comfortable enough with my gut check to challenge that baseline, so I started down a path of figuring out why my assumption is wrong. I would rather start off defeating my own thought with data before going down rabbit holes.

So here is what I did:

Go to one of these tools, look for your own domain info and look at what they say you daily ad budget is.
Copy that daily ad budget and multiply it by 30 (this is what the tool says your monthly spend is)
Go to adwords and choose the last 30 days (if you have a huge up and down on spend, get a longer time frame and determine a consistent 30 day average)
Figure out what the observed percentage difference between the tools number and the actual.
Once you get that difference add it to the baseline in the tool. (it should be the same as your adwords account now.
That baseline percentage difference (in most cases) applies to all of your competitor’s accounts. You can get a better feel for their spend (not 100% accurate, but closer than what these tools say)

So after I ran the number that way, I felt even more so that my gut was closer than the tool, so I need to validate my data. So I hopped on live chat, asked the rep at one of these tool vendors why the number was so far off from reality and how its derived. The rep explained that with these tools they and most companies will eliminate multiple keywords from campaigns and ad groups and treat them as a single keyword.

What this means is that if you have potted plant, potted planters and potted soil enhancer they will likely remove two keywords and use just one. I get that, but depending on how campaigns are set up, this can eliminate hundred, thousands, or even more keywords from the daily ad budget and provide a misrepresentative number.

This is AWESOME news for some people because after you do the math, the formula may not apply to everyone the same way. As such the result will be wrong data all together.

The end result of this is that PPC research tools are good for research, but trying to use them as a source of accuracy and planning needs to be considered strongly before actually using any of the data,

Even the keyword lists these tools provide are very broad search query reports and drags in everything that served up an ad on a domain, so it does not mean that you or the competition actively bid on these keywords.

By mgoldberg | Posted in Marketing | Tagged BI Tool, Google AdWords, Keyword Spy, Marketing, PPC, SEM, SEM Rush, SpyFu | Comments (0)

Email Automation – When It Goes Wrong

January 20, 2013 – 7:27 AM

You may have read some of my other articles regarding email automation. I actually like using automated tools for email because it keeps the train running on time and if set up correctly works very well for the masses.

However, some email campaigns are somewhat complicated in how they are scheduled, connected to lists, and tied to other databases. Because of these occasional complexities in email campaigns can really affect revenue if not run optimally, Automation might actually hurt the process because the thing about automation is that it is kind of like the old Ronco Rotisserie “set it and forget it”. Auto pilot works only when it is behaving as normal, but if its not working, its only going to get worse if you are not looking at the data correctly.

I took over an email campaign internally a month or two ago and while I could have used an automation tool for these emails to be sent, I elected to use Gammadyne. You can read more about this handy tool here: http://michaelagoldberg.com/?p=42

This particular campaign two years ago was showing revenue growth month over month year over year on average of 150%. A lot of this had to do with the volume sold the previous year, so obviously that will affect the numbers. However, in this particular campaign, we went from 150% YoY to flat to negative percentages in rev last year. This kind of revenue stream is not supposed to decline for years (when doing some math it would literally take years to start declines unless you sell NOTHING the prior year. If you have not noticed, this email campaign is intended for reoccurring revenue from prior years from prior customers.

So when I took over this function, I tried to use the automation tools we have for email and the results were only slightly better. A lot of this was out of my hands at the time, but the decision was made for me to take 100% ownership of the campaign.

When I did, the first decision I made was to revert back to Gammadyne (seriously, read the article about this) for sending emails. There was a glitch in the automated tool that caused complication with the amount of touches per email address, and also if they were scheduled too close, they would overwrite the current list that was sending emails. So on my mind the reversion was an absolutely necessary function.

Email copy never changed, but I did modify the way we pull a list. I made it less technical and just simplified the overall strategy with the mindset of “what methods will product the most revenue?”

Execution went as planned, we scheduled the emails to go out at a specific time optimal to customers reading email, and process management was enforced. Meaning while automation has the benefit of “set it and forget it” it can take the intensity and focus off of a project sometimes because you are under an assumption that everything is fine. I think human nature sometimes plays a part subconsciously because if its out of site and out of mind, its OK to many people. I get that, I never said I was not critical about execution J.

So my thoughts on this campaign were simple, before the campaign started my primary objective is to stop the bleeding of revenue. That’s the only goal I cared about. With that said, the team knew what the plan was, how it was to be executed, when it was to go live, and what the goal was two weeks before launch. Setting proper expectations from day one is the single most important priority for me, this way no stone is unturned.

So we had all the process set up accordingly, we send at X time, every week, keep a clean list so we do not get flagged for spam, and watch read rates and revenue. Because YoY revenue would not be a fair gauge to the improvements I decided that Month vs. Pervious month was the best way to get a good read on the effectiveness of the email campaign. Revenue was the only thing I cared about. Below are the results and please note that December is typically the worst month of the year for consumer as our industry is not a Christmas gift.

Dec vs Nov = 36% increase in revenue the first month we took this campaign over!
Jan vs December = 71% increase in revenue in month two and the month is not complete!
Jan vs Nov = 82% increase in revenue in less than 60 days!

People who know data, are saying HOLY COW that’s a major jump. To make it even more appealing, this is not chump change, we are talking about a lot of money.

So in this particular instance, a manual non-automated process for email campaigns out performs in a major way. This is not a knock on automation its more so to demonstrate that automation has a place in marketing, but so does non-automated campaigns.

By mgoldberg | Posted in Marketing | Tagged Eloqua, Email Campaigns, Email Marketing, Gammadyne Mailer, Marketing Automation | Comments (0)

Google AdWords Tip: Keyword Match Types

December 27, 2012 – 8:55 PM

I am going to say it straight away, if you are not using more than one match type per keyword to control your ad spend, you are making a mistake. I think anyone who happens across this blog knows a thing or two about PPC, or perhaps they are doing some kind of research to better understand account behaviors.

Let’s just get to the meat of this one:

Example: You are managing one campaign in your account and in this campaign you decided that you wanted to be granular enough to create one ad group per keyword set. Notice I said keyword set because one thing I have seen in many accounts is that most people go one of 4 ways on keyword bidding:

1) broad match terms

2) broad match modifier

3) phrase match terms

4) exact match terms

There is not a single thing wrong with any of these types (except for the modifier, I do not like to use them at all), but how are you going to bid on these keywords? And more importantly how much do you think you want to spend on them? Even more important, how will they convert?

Simply put, you are cheating yourself by bidding on just one of the match types because one of two things will happen:

1) You will have a very broad funnel and spend too much on unwanted traffic.

2) You will have a very vertical funnel and not have as much traffic as possible.

There are always going to be winners and losers, but you can control the risk by doing the following:

Example: Keyword is airplane tickets

Airplane tickets – Bid = 3.00
[Airplane Tickets] – Bid = 7.00

You can always add other match types to the ad group, but by doing this simple albeit redundant keyword set, you will still show up for things like “discount airline tickets”, “plane tickets”, “airfare” etc., but spend far less doing it. And with the money you just saved you can jack up your exact match bids for higher conversion and click thru rates.

The only down side to this kind of bidding strategy is that you must make sure that when you are adjusting bids, you don’t lower the exact match below the broad or you will pay more in the long run.

By mgoldberg | Posted in Marketing | Tagged Analysis, Google AdWords, Google Analytics, Lead Generation, PPC, Qualified, Reporting | Comments (0)

Social Networking Pro’s and Con’s

December 27, 2012 – 8:30 PM

I think that SM has definite value as an additive to an overall marketing plan, but it’s not been incubated long enough to create a solid revenue stream for the vertical I am involved in.

I have a love hate relationship with social networking in general. As a branding tool, it can provide your name to an audience over time that can eventually reap benefits, but if you think about banking social network traffic to a sale, you might not be happy with the results.

Here are 5 things I like about social networks

1) A great tool to communicate with people that have an interest in your vertical.

2) Branding to a large audience for almost next to nothing is a refreshing change from search.

3) You can speak “loosely” without having to worry about the typo police to come after you.

4) The level of engagement you can have with people you connect with is much more focused and personalized than anything an article or web page can provide.

5) You can expand your audience literally with the click of a mouse. You do not even need to say a word either.

Now, here are 5 things I do not like about social networking:

1) Some verticals really have no place going social. Seriously, is there any social value to this page?

2) There are no take backs on posts, if you have it wrong, you will suffer the fate of the masses.

3) It takes a long time to get a following.

4) Unless you are a social rock star, traffic is weak at best

5) How much money are you really making from social marketing?

I am sure that some people have been very successful with social marketing, but I have been around this block a few times over the years and have personally had the most lift and success with social marketing when I pay for ads to whatever page I have. This worked out well on Facebook and LinkedIn. Twitter on the other hand is actually pretty cool and you can get similar traffic, if not better traffic for twitter than FB or LinkedIn.

I did not say too much about Google+ because while I have been using it for a couple of years now (since beta) I do not really dedicate a lot of time to it. Well, actually recently I have (with the help of an excellent graphic designer) started the process of building out a solid Google+ business page.

By mgoldberg | Posted in Marketing | Tagged Facebook, In-Stream Ads, Mobile Site, Segments, social networking | Comments (0)

6 Reasons to Look Beyond Google AdWords Conversion Data

December 19, 2012 – 7:30 AM

More often than not, I read an article about PPC and the biggest item on anybody’s mind is return on investment (ROI). Why is ROI such a big focus with PPC? Mostly because it is a major metric to gauge the effectiveness of any campaign that you are running. Even I look at this metric to make day-to-day decisions in accounts.

However, I cannot emphasize enough that it is a single metric, not the only metric in the world to live and die by. Below are a few reasons / items to consider looking beyond spend vs conversions.

1) Focusing solely on ROI will minimize your audience. I have seen accounts that ran on manual bids and when the lever had been pulled to focus on conversions / ROI, a 45% decline in impressions and 50% decline in clicks occurred. This could have been avoided by watching the account closely and not focusing on just spend vs. conversions.

2) Killing zero conversion keywords without proper context is like playing Russian roulette. Not all keywords convert the same, but not all zero converting keywords mean they do not participate in a sale. *Hint look in to the accounts search funnels. If you are focused on Sales data outside of AdWords, you can also reference increases or declines in when adjusting keyword status and bids.

3) Let the LANDING PAGE do the converting. I understand that AdWords needs to play a big role in conversions, but in a lot of accounts you work with are only going to convert as well as the landing pages content, layout and calls to action are optimized. Getting it right in AdWords is only part of the battle. A poorly constructed landing page will almost certainly wreak havoc on your campaigns and make all of the keywords pointed to it look like poor converters.

4) If you have access to sales reps, ask them how the calls have been when making changes to the AdWords Account. If you are working an account that has a number to call, know that there will be untracked sales that are directly linked to your efforts. Once again, you would likely never even know this metric existed if you were focusing on spend vs conversions.

5) Do not strangle your account with profitable negative keywords. Add negatives for one reason only, because it is an irrelevant keyword. Don’t try to be smarmy and add negatives to force ad serving on specific campaigns. The intent sounds good, but why would you take any chance at all on profit?

6) Go with your gut ONLY if you trust it. I am the first person that will admit that I take risks, but those risks are calculated and also based on a gut check I trust which is me + others input. I can only say that if you are not comfortable with your own gut check, do yourself a huge favor, find someone who knows the account, or sales flow and ask them. More often than not, you could build your own gut check by learning from others you trust. Looking at conversion data will never allow you to have a fully developed gut check.

As you can see, ROI is not the only metric in the world to look at, it is one of many metrics that provide you with the data in a sale. There is a lot of data in the AdWords Account that go beyond the standard ROI metric that you really need to consider looking at. There are also non account related metrics that might be harder to attain, but regardless, should be explored if you have access to that data as well.

By mgoldberg | Posted in Marketing | Tagged Analysis, Google AdWords, PPC | Comments (0)

Inbound Sales Tip – Queue Based Calls Create Mediocrity

December 19, 2012 – 6:17 AM

If you have a small team of sales reps the following message is for you. If there is one thing that I have learned about inbound calls over the years it is that sending calls to a queue is a bad idea.

Sure you are being fair and just by creating an even playing field for all sales reps to get their fair share of the money. They will each have their turn at bat through out the day to hit that home run.

But what you might be doing is sacrificing revenue for fairness.

How fair is it to the company to put your top performing sales reps in a position to make less money than their potential so the sales reps with the weakest numbers gets their fair share of the calls?

Sales people are supposed to be competitive, not complacent. I am sure there are a bunch of different sales strategies, but my observations show me that if you give the sales calls to the hungry rep, the company will be rewarded.

Isn’t this a DE motivator because you are not permitting the weaker performer a chance to make money?

This is SALES, not lets be friendly and make nice nice J. Sales reps know the deal, sell or suffer the consequences. If your single task is to make money and you are not doing an optimal job at it, why should a company, a team, and even a coworker suffer because someone does not take advantage of every opportunity they get.

Queue based calls might work for big companies that deal with huge volumes of inbound calls, but if you have a very transactional sales force and can map out the days where someone has their turn at bat to get the lion share of sales calls, give it a shot because it creates a more competitive environment and should increase sales across the board.

By mgoldberg | Posted in Sales | Tagged Phone Sales, Sales Tip | Comments (0)