I recently passed the one year mark at my current startup. As I looked back at the things we’ve been able to do over the past year, and compared and contrasted that with the first year’s goals and efforts at my other two startups and those of my friends running marketing elsewhere, it occurred to me that there’s a great deal of overlap (probably because there’s a lot of commonality among our goals, efforts, and achievements).
So, I figured I’d make a checklist of sorts – the things one has hopefully accomplished by the end of the startup’s first year in the marketplace. To some extent it’s a follow-up to my earlier eBook, “Building the Marketing Plan: A Blueprint for Startups.” Though this time around, it’s less about putting together a plan, and more about proposing a specific criteria for evaluating progress during, and at the end of, the first year.
This is by no means a definitive guide or a comprehensive list, but rather a proposal for what might be reasonable for a startup to accomplish, and questions you probably want to answer, within the first year. (And of course, depending on your specific situation, some of these may be much harder/easier, and more/less important than others.)
So, here's the full eBook, as well as a summary checklist below:
(The eBook covers each of the items below.)
Refining the Web site
- Identify key personas
- Develop content for each persona
- Develop content for each stage of sales process
- Rich set of calls-to-action
- Knowing which pages are popular, which are not, and why
- Frequently measuring, tweaking, experimenting with content on your site (not only the home page – but also the calls-to-action, organization of the pages, layout)
- Clear, compelling to all prospect personas
- Differentiated in the market
- Pressure-tested in sales, in PR
- An active blog
- Webinars, videos, presentations, eBooks,
- Lots of content addressing your prospects’ problems
- Know which content is popular
- A rich pipeline of content in development
- Content for each stage of sales process
- Know which terms
- Prioritize based on popularity and difficulty
- Build content targeting these terms
Case studies and References
- Rich case study library for key markets, use cases, hard- and soft-metrics
- Leverage in PR, web site, videos, webinars, live talks, reference calls
- Identify key events
- Exhibit and present at events
- Measure ROI
Which social media sites matter
- Identify key sites
- Measure importance LinkedIn, Facebook, Twitter, etc.
- Understand conversion rates (traffic==>leads==>customers)
Connection with bloggers
- Know the key bloggers
- Engage with them as part of a community outreach program
- Pilots to measure lead gen options
- Understand cost/lead and cost/sale metrics
- Know (and deliver) lead flow rate required by sales
- Conversion rates for traffic, leads, customers
- Understand how competitors compare with respect to traffic, inbound links, offers
Lead nurturing, lead intelligence
- Understand buying process from customer’s perspective
- Build content/tools/process that naturally pulls prospect along the sales stages
- Leverage automation to streamline nurturing
- Identify awards that matter (for companies, products, individuals, campaigns, etc.)
- Develop your award program
- Pitch, learn, refine
Mapping the Sales process
- Know the buying process
- Deliver content, tools, workflow, automation to support sales process
Prospecting tools and selection criteria
- What are the key variables to use? (prospect title, company demographics, etc.)
- Design an experiment to zero in on selection criteria?
Sales rep training
- Develop onboarding process for new sales hires
- Know sales cycle, conversion rates at each stage
- Know average sales price (ASP) and cost of customer acquisition (CoCA)
- Know key productivity metrics per sales rep
Leveraging lead intelligence in Sales
- Know prospects’ behavior on your site
- Append demographic data from external sources
- Automate lead grading and nurturing using available data
Iterate & refine
- Set Goals
- Web site content, layout, navigation
- Calls to
- Sales/marketing infrastructure
Know where to step on gas
- Lead generation programs
- Cold-call list selection criteria
- What roles to hire next
- Marketing automation
- Lead nurturing
- Competitive intelligence
Thoughts on other stuff to add to the "must achieve" checklist?
Here is my guest post on the HubSpot blog about how journalist Darren Garnick and his 9-year old son Ari did some very cool "newsjacking" recently. (According to marketing strategist David Meerman Scott, 'newsjacking' is "the process by which you inject your ideas or angles into breaking news, in real time, in order to generate media coverage for yourself or your business.")
Outlined in the post are several marketing lessons in both how the candidates answered, and in the propagation of the story itself, including:
And, when you're ready to newsjack, there's some great tips from David Meerman Scott.
Inside sales teams live and die by the quality of the leads they call on. Though you hopefully you have an ever-growing amount of marketing-generated leads, chances are that if you’re in a growing business, it’ll feel like a never-ending treadmill, and you’ll be adding inside sales professionals at a faster rate than marketing can generate leads. (Yes, there are the rare and notable exceptions – companies like HubSpot that have insane inbound lead volumes, or companies totally committed to the “freemium” model where inside sales reps focus 100% on upgrading downloaders of their free products.)
The economics of “high velocity” (a.k.a. “low friction”) sales models are particularly sensitive to the quality of cold-call lists, for the following reasons:
- The average sales deal is relatively small
- Sales expense makes up the majority of the cost of customer acquisition (CoCA)
- With cold-calling – even targeted cold-calling – you’re typically looking for a needle in a haystack, with somewhere over 99% of the calls not contributing to a sale.
As a result, success rests on that razor-thin difference in quality: a list that’s “99% bad” might drive profitable sales, whereas “99.9% bad” might be a giant waste of time, because the former has 1% quality leads, whereas the latter has 0.1%.
In this post, I’ll outline a technique for systematically zeroing in on the correct selection criteria.
Selection Criteria: Identify the variables
Whether you’re using list brokers, online sources like Hoovers or iSell, or online directories such as Jigsaw, ZoomInfo, and LinkedIn, there’s many sources for finding prospects to call on. They all have pros and cons, but regardless of the source, everything starts from the selection criteria you use. Though the exact criteria is going to be specific to your business, several obvious variables to consider are identified below. And for each factor, identify the different levels you want to explore. (After all, you might not yet know the right title, or the right company size.) In the example below, let’s assume we have a software product that helps marketers, but we’re not yet sure the right title to call on, or the right company demographics.
- Prospect’s job title (and if there’s several common synonyms, you might consider grouping them all using a keyword in job title, job description, and/or experience). For example: “marketing communications,” “product manager” or “public relations.”
- Prospect’s level in the organization – is this an individual contributor? A manager/director? A C-level executive? Many databases either provide this automatically, OR you can fairly easily group them in Excel, by looking for words like “manager” or “director”, or “vice president” and “chief”.
- Department: depending on whom you sell to, you might pick several groups (marketing, information technology, customer service)
- Company size (Revenue and/or # of Employees)
- Industry sector (e.g., healthcare, financial services, manufacturing, etc.)
Here, you might want to be careful just how many different levels for each variable you choose to test. The more variables you have, with more levels, the larger the sample size you’ll need for the experiment (see below).
What to Measure
Though you might want to measure a bunch of factors, three typical ones to consider are:
- Conversion rate (to next stage in the sales process – e.g., a demo, or a “needs analysis” call)
- Deal close rate
- Average sales price
The benefit of measuring #1 is because it’s the quickest statistic you can get, as compared to #2 and #3, which require going through the full sales cycle (which not only results in a delay, but you may also need to have a sufficiently large sample size upfront, because each stage in the sales process is a reduction filter). Though over time you’ll of course want to also measure the close rate and deal size, as that is the ultimate arbiter of quality and profitability.
Establish your baseline
If you already know what your average conversion rate is, or deal close rate is, great. Unless you’re just starting out, you should have a decent sense of what those are. And if not, it should be straightforward to calculate. Just be careful of the sample bias: if a particular criteria dominates your sample size (e.g., you’ve primarily called on marketing communications professionals at healthcare companies with revenues between $50M and $250M) then your “baseline” might not be representative of the general population. But the good news is, however biased your baseline might be to begin with, once you’ve designed the right experiment (below), you’ll definitely have a reliable average after you carry out the experiment.
Design your experiment
How many calls do you need to make? How many samples do you need for each combination of variables? This is an important question, and not one that can be answered trivially. It depends on many factors, including:
- How many variables are you testing?
- How many levels for each variable?
- What data is easily available from whatever database(s) you’re using?
- How many sales reps do you have making calls?
- What are the typical conversion rates?
- How much precision and statistical significance are you looking for?
- Is it ok to assume the factors are all independent?
I’ll skip the rigorous statistics, and instead propose a couple rules of thumb:
- If you’re measuring something that happens infrequently (e.g., if the 1% of leads that convert to demos), then you might want to have at least several hundred names for each combination of variables. After all, if you’re trying to see if there’s a difference between variable combination A and variable combination B, you need a large enough statistical sample to see a difference between, say, 0.8% and 1.5%.
- You don’t need to test every permutation. For example: let’s say you had 3 different titles, 5 different industries, 4 different company size categories, and 5 different industries – that would be 300 different permutations. If you populated each one with, say, 300 names, that would be 90,000 names in your experiment! (First, good luck getting that list with all the variables, and second – that might keep your team busy for the next 6 months.) The good news is that if you can make some simplifying assumptions (e.g., that the factors are largely independent), you can use something called fractional factorial design or orthogonal arrays to dramatically reduce the needed sample.
But you don’t need to be a statistics wiz to carry out this experiment. At the end of the day, just try to have at least a couple thousand names to call, and you might be in fine shape to get good insight into which factors matter. After all, this is less about proving something in court, and more about quickly iterating to identify a good list selection criteria.
Execute the experiment
The neat thing with high velocity models is that you can do these sorts of experiments fairly quickly. An inside sales rep should be able to make 100 calls per day. If you have a team of five reps, that’s 2,500 calls your team will make this a week. So within a week, a picture should start to come into view.
Analyze the data
As you carry out the experiment, compute the results along the way. Do a sanity test and ask yourself a few questions such as:
- Do the results make sense? Are they statistically significant?
- Is there much variation across sales reps?
- Based on the results so far, what might be the best selection criteria?
Along the way, you might decide to do more experimentation, further refinement, or to beef up the sample size. Here’s a typical example from a cold-calling experiment, with some notes regarding each variable tested:
- Prospect title: looks like the “marketing communications” title performed best, with calls converting to demos almost twice as good as the baseline (2.2% vs 1.3%).
- Department: looks like calling on prospects from Marketing was more productive than calling on IT or Customer Service teams.
- Level in organization: in this example, we are looking at the conversion across the entire sales funnel, from the initial call to the sale. Note the wide variation here: as we move from individual contributor to managers/directors, the conversion rate rises by almost 3.5X (from 0.20% to 0.68%), then falls off sharply as we call even higher up the organizational structure. This might be the case because you might need to reach out to someone with budget authority; but as we call higher in an organization, we might reach someone with significantly broader responsibility, who is not looking closely enough at the kinds of challenges our software might be solving. Of course, there’s other reasons why there might be a sharp drop-off as we call higher, so it’s possible you might want to do further investigation. (For example, your connect rates might be a lot lower once you reach a management level that relies on administrative assistants and voicemail to screen calls. Or, your sales pitch might not be tuned to the needs of the senior exec.)
- Company size: So given the results in this example, what companies should you target? My sense is, it depends – on several factors that we’d probably want to look closer at. On the one hand, it looks like as the company size increases, our average deal size increases. On the other hand, we have not looked at other factors, including: How many calls (and related sales effort) did we have to make for each sale? How many companies are there in each category? Is our entire sales team capable of selling to the larger firms, or are just the best few reps able to make these sorts of sales? Also, we can see that the ASP goes up for the largest firms, but not by much – so we’d probably want to pressure-test whether this is truly representative of the opportunity within larger firms, or just a reflection of our sales capability to sell to major accounts. Depending on the answers, we might, for example, conclude that our sweet spot might be in the $50M-$250M company size – a large market segment, with a good ASP.
I should point out that there's a bunch of simplifying assumptions here. For example, that the variables are all independent – in reality, it may turn out that some combinations are much rarer than others, and you can’t source them as easily. Or that other factors that we might not be testing for – e.g., list accuracy – also matter. Or that rather than there being one ideal target, there in fact may be several demographic profiles, or combination of variables, that represent great targets (cluster analysis might be a way to identify these – perhaps a topic for a future post). But hopefully, over time, you can identify the variables and values that yield the best results, and identify more variables to test. After all, the inside sales model is a fantastic laboratory for rapidly testing, refining, and improving not just your sales operation, but the entire customer life cycle - how you market, sell, and support customers.
Finally: I’d be glad to hear from you – which factors have you found particularly valuable when building cold-call lists?
Recently, I needed to quickly build a geographic opportunity model. I've written in the past about some ideas for market sizing, and a couple of those ideas came in really handy - so I figured I'll outline the approach and provide a hands-on example.
In this particular case, we needed to load-balance a sales team across geographies, and wanted to get a sense of how to split up the territories, and get some insights into how to allocate sales efforts within those territories. But the same sort of analysis can be applied to many other common questions:
- How is our penetration in territory X compare to territory Y?
- How is our success in industry X compare to industry Y?
- How many prospects should we expect in territory X?
- How many lead generation events should we do in California, versus New York, vs Connecticut?
Now, in the ideal world, you would have access to a database of the exact thing you need - for example, "companies in industry X, having revenues between $200M and $1B, with data centers of at least 50 servers." But in the real world, you might not have exactly that. In fact, you might not yet know the exact target demographic you even want. So what to do?
Well, the good news is that it (almost) doesn't matter the exact metric to use! (This is related to the Law of Large Numbers and the Central Limit Theorem, though I won't get into that unless there's a comment asking for it!) As I outline in more detail here, there's a multitude of free sources online - by industry, by state, by postal code, by products produced, by company size, etc. (For example, here's info on gas stations across the U.S.)
To illustrate this point, let's take a specific example. Let's say you're an enterprise software vendor, targeting a particular role within organizations above $100M in size. And let's say you want to carve up the U.S. into 5 contiguous territories, with roughly the same opportunity - how might you do this, and how much confidence should you have in the model?
So consider 3 metrics:
- # of Companies above $100M in revenue
- # of Coffee shops
- # of Pet Grooming salons
What? I'm proposing sizing an enterprise software vendor's market opportunity by looking at the # of pet grooming salons? YES! The cool thing is that there's often a great correlation among seemingly independent quantities. So if you didn't have one metric, and you pick a sufficiently close one, you'll be in good shape -- because even seemingly UNRELATED metrics are often highly correlated. Let's take a look. Consider the following - REAL - data of these 3 metrics, by state:
Turns out statistically, they are indeed highly correlated:
Another Example: Correlating Visitors, Leads, and GDP
As another real-life example, consider the following 3 metrics for a start-up company, with the key question being, "what is the relative market opportunity for each state?":
- # of Website visitors from each state
- # of leads from each state, generated via marketing activities
- the GDP (gross domestic product) of each state
The interesting thing is that though this was a start-up with just a few months in the market, all 3 metrics are quite highly correlated. If they weren't correlated, then we might wonder whether one of the metrics might over- or under-represent the opportunity. But they are correlated - so using any one of them - or better yet, averaging several metrics - would likely give a reasonably accurate measure of the state's opportunity.
And you don't even need data across all 50 states (or all industrialized countries). You can use this technique when you have just a couple data points. For example, let's say that you want to roughly size the opportunity in Industry X (a market you have not yet entered), to see how it might be compared to Industry Y -- where you may already have a presence. So find some metric that might roughly correlate with opportunity (e.g., annual revenues, or # of firms) of the two markets, and their ratio will give you a sense for how you might do in the new segment.
So next time some asks you to size a territory's opportunity, feel free to ask, "Sure - how many dog grooming shops does it have?" :)
Last week, I was on a fun panel at the Enterprise 2.0 Conference called " Can Inbound and Outbound Marketing Co-Exist?". (Moderated by Brent Leary, the panel also included Michelle Burtchell from Constant Contact and Mike Lewis from Awareness.) Summarized below are my answers to some of the questions.
Definition of Inbound Marketing
Turning your website into a magnet for prospects, ways to get them into the funnel, move them along the sales funnel – and accomplishing all this through content (and amplified via social media and SEO). (Largely plagiarized from HubSpot.)
Compare and Contrast IM vs. OM
- Inbound: low cost per lead; takes time to get a serious volume of traffic, subscribers, leads; requires commitment and investment by team to create content; but is an “annuity” – traffic and leads that continue to flow;
- Outbound: faster “0 to 60” to get going, higher cost per lead, lower conversion rates, better for targeting accounts;
Does Inbound Replace, or Enhance Outbound?
Inbound is a huge enhancement to outbound, but you (probably) need both. Depending on the business model, dial can be at 80/20 or 20/80 (and anything in between). I say “probably” because there’s certainly some freemium models that generate enough inbound leads to not require much outbound activity; or the cost-per-lead constraints are such that outbound leads might not be profitable.
Currently, for us ~80% of our sales reps’ activity is taken up by outbound (a dozen reps need a lot of leads to fill up their day!) Though we’re seeing the inbound/outbound activity ratio move up as more inbound leads come in). And interestingly, >50% of the business is coming from inbound marketing – those are inherently better leads.
Compare Inbound Marketing at a Large Enterprise vs. SMB
- Enterprise Marketing Team: With deeper pockets, you can have a bigger team, with more specialization. Some might be focused on blogging; others on video production; others on creating white papers / eBooks; events; etc. With a larger marketing team, you have more of an established base (of content, of programs, of knowledge, etc.) - at the same time, there’s more inertia.
- At a smaller company or start-up, there’s less of a foundation, you need to learn quickly and refine. But the good news is that you can, and it’s easier than at a large firm. Though it’s all the more important to have the right team members – can’t afford to make mistakes when you only have 2-3 people in Marketing. And, they need to be comfortable wearing multiple hats.
- But, whether at a large or small company - basics are the same: figure out what the learning and purchasing process is from the customer’s perspective, and design an experience to facilitate that.
How Does Inbound Change Marketing Team’s Make-up?
- “Traditional” marketing teams: activities and deliverables included direct mail, events, collateral. As a result, skills were around copywriting, events, creative messaging pros, collateral production, direct marketing campaigns;
- “Inbound” marketing teams: Because everything starts from useful content, first and foremost you need content creators – passionate domain experts that can help educate your prospects. Once you have the content, you’re ready to promote it via social media, so you need evangelists and networkers – folks that can do their magic both in person (cocktail parties and tradeshows) and online (Twitter, LinkedIn, etc.). And finally, you need analytical optimization-focused “geeks” – to understand where success is coming from (so you can invest more there), and what to tweak (to improve marketing ROI). Here's a great post on this from the HubSpot team.
How Does Inbound Impact Customer Relationship Management?
Increasingly, the prospect/customer line is blurred. For example, for “freemium” products and inbound marketing : there’s a natural progression from trying the free product, to becoming a prospect, then making a 1st purchase order, a 2nd purchase order, etc. With our software, folks running a data center first download the Community Edition, then purchase our premium product for a single cluster, and 3-6 months later may purchase it to help manage the entire data center. The marketing team is responsible for delivering a smooth nurturing process that takes the customer along this journey.
Impact of IM on Marketing/Sales Relationship
Sales and Marketing should hopefully be joined at the hip – this goes without saying. For example:
- Conduct regular reviews of key metrics – activities (leads, calls, demos, etc.) by lead sources (various inbound programs, events, tradeshows, etc. as well as outbound cold-call lists, etc.) – and results (opportunities, deals)
- Jointly develop a sales/marketing “service level agreement.” What constitutes a “sales-qualified lead” that Marketing endeavors to deliver (e.g., “someone who download a product trial, attended a webinar, and watched 3 online videos)”, and what Sales will do with it (call within 1 business day). After all, some leads are much hotter than others, in terms of both conversion rate and half-life.
- Define a nurturing process: there’s several key groups that all need to be nurtured in their own way – customers, prospects, leads, and cold-call suspects. So working with Sales, Marketing ought to define a “curriculum” that – over the course of 6+ months and multiple touch points, the prospect is touched via some activity, and is presented with some relevant content, with an opportunity to take an action along the way (attend an event, a webinar, download an eBook, see a demo, etc.)
If you’re not doing inbound marketing, you’re trying to clap with one hand. Or put another way, you’re missing easily half the opportunities out there, and are bearing a much higher cost per sale than you need to be. If half your leads are NOT from inbound marketing activities – it’s a missed opportunity. If you’re not nurturing customers throughout the sales cycle – that’s lost productivity, and represents extra (and expensive) work that your Sales team winds up taking on.
- Content creation is key to drive inbound marketing. Use the blog as a “content laboratory” – see what people are interested in, what drives traffic, comments, interest. Then build out the most interesting content in variety of ways – eBooks, webinars, videos, tutorials, etc.
- Integration of sales/marketing systems is key – need to engineer this just like you engineer your product.
- “Thermostats, not thermometers” – don’t measure things just to know the metrics; do something with the data – take action, adjust something about your process, the campaigns, the business.
So I finally decide to build this little site (ilyamirman.com), primarily to serve as a parking lot for my hobbies (photography, comedy, politics, marketing). To build it, I looked at a bunch of alternatives (Wordpress, Drupal, HubSpot, others) and in the end settled on HubSpot. One key requirement is that because this is a hobby site, during the week I can't spend more than few minutes on it, so it had to have a very low barrier to entry, and easy maintenance.
I've been a HubSpot fan (the product and company) for a few years now, and was once again reminded why:
- FAST and EASY: I am no web programmer, and don't register domains or set up sites frequently. But when I decided on a whim to finally do this, it was literally just a couple hours between grabbing the domain at GoDaddy and building the site in HubSpot on a Saturday morning.
- CMS: The HubSpot content management system is actually fun to use. It's a pretty effortless way to create pages, with lots of options for page layout and formatting controls. And as powerful/flexible as they are, I've found the user interface more intuitive than my experiences with WordPress and Drupal.
- Hosting: I like the fact that I don't need to worry about where the site lives, uploading updates, managing the server.
- Blogging tools: It's super-easy to set up a blog (or several blogs), and insert related modules (recent posts, tags, etc.) throughout the site. Furthermore the authoring tool includes automatic suggestions/notifications, such as "article body missing images" and "meta description missing keywords" to help optimize the blog for readers and search engines. And, HubSpot takes care of all the "plumbing" - feedburner, social media sharing buttons, etc.
- Analytics: I used to be a Google Analytics junkie, but find HubSpot's analytics both easier to work with, and more actionable and informative. Whether it's looking at traffic sources and volume, page popularity or dynamics of keyword rankings, blog analytics or clickthrough rates on lead follow-up emails, they're all there, just a click away.
- Forms:It is so easy to add all sorts of forms, without hooking up your own back end for the leads database.
- Easily insert widgets: Whether it's adding Twitter feeds, photo galleries from my SmugMug photo site, SlideShare presentations, or a ton of other widgets - it's easy to pimp out your HubSpot-based site.
- And the biggie: INTEGRATION! All the tools are nicely connected, woven together and talk to each other. For example:
- Once you enter your social media accounts (Twitter, Facebook, etc.), new blog posts trigger notifications on those accounts (rather than you having to post to Twitter etc. manually).
- Once you enter the keywords relevant to your business (or pick them from the keyword tool), it will automatically notify you of conversations/posts/tweets/etc going on RIGHT NOW on those topics, so you can participate in the blogosphere.
- The forms, leads database, email tools and lead nurturing are all integrated, eliminating the need to string together several plug-ins, databases, and workflows.
- Content creation - whether on a web site page, a landing page, a blog post, lead nurturing emails, etc. - features a consistent UI.
- HubSpot's analytics are monitoring all aspects of the site, everything from what drive traffic, to what your visitors are clicking on, what's generating leads, nurturing campaign clickthrough rates, etc.
- Tools like Page Grader or Link Grader provide actionable intelligence on what to improve, and you're just a click away from going right to what needs fixing.
(And if I was a business, I'd also care about HubSpot's other features, such as integrated email and lead nurturing tools, integration with Salesforce.com and other CRM tools, landing pages, all sorts of analytics, etc...)
What to Improve
There's a couple things I'd like to see improved:
- CMS: In the CMS interface, sometimes there's weird behavior, like when you delete something from the middle of the content module, it pops you to the top of the module (forcing you to scroll down). There's a couple quirks like that with bullets as well.
- CMS: Cutting/pasting from MS Word - it tends to paste in formatting that might screw up some browsers. Though the CMS includes a "remove formatting" button, you need to remember to click it. I think this used to be part of the cut/paste interface and gave you the option automatically.
- Trial: If you just grabbed a new URL (and it's therefore not a live site), HubSpot's trial software might not allow it, forcing you to temporarily choose a different name. But, elements of that "temporary" choice persist, though hidden from your site visitors. (For example, you'll get some site notifications from that original domain name.) Better guidance on that, or eliminating that constraint, would be great.
Sure, you can weave together half a dozen technologies – content management, analytics, hosting, SEO tools, campaign management, etc.– and with a bit of effort get it all to work. But, you risk spending too much time connecting disparate systems, customizing, and inevitably running into limitations – and every second of that will take away from doing what you actually want to do. And so the bottom line with HubSpot is that there's now no barrier between content creation and content publishing.
Last week a friend asked me about market sizing methodologies, and I dug out this presentation from a long time ago. Back in 2003, I did a talk for the Boston Product Management Association on the topic of market sizing, a common marketing challenge. Specifically, on a bunch of techniques to size existing markets, for both top-down rough estimates, as well as territory analysis and setting sales quotas.
There's no one source of information that's totally reliable and comprehensive, so what I've found helpful is to use multiple approaches to zero in on an estimate. The good news is that metrics that correlate with product usage are widely available in many sources, include government databases, relevant publications' reader studies, and a variety of commercial databases.
Following are the steps I've found useful in building the Total Addressable Market (TAM) model:
- Find a widely-available metric that correlates with demand for your product;
- Carry out statistical analysis to identify the scaling factor between demand and this metric;
- Get database with the metric by industry and geography
- Build the model;
- Verify model’s predictive power.
Here are the slides (heads-up: some of the links from 2003 might not work). (Also - here are the two magazine circulation audits I mention on slides 24-26: Computer World
and Internet World
Last summer I wrote a little eBook for building a start-up marketing plan, and published it on HubSpot's marketing blog. (I was pretty surprised that >40,000 folks downloaded it.)
Here's the eBook:
Last year, HubSpot's Mike Volpe and I spoke at a conference in Chicago about lead generation using inbound marketing. Here's the presentation, explaining the ideas behind inbound marketing, and using my last start-up (Cilk Arts, acquired by Intel) as a case study.
In a nutshell, we at Cilk Arts used our blog as a "content lab" - quickly publishing a broad range of content, pressure-testing what's of interest to our audience, and developing deeper content in those areas. We then repurposed the content (blog posts, tutorials, web videos and demos, eBooks, etc.), used organic SEO to get found, and used social media to let more people find out about us. As a result, without spending a dime on outbound marketing, over 100,000 developers found out about us. (Check out the slides for more details...)