7 Ways to Make Twitter an Effective Channel

Twitter can be a very effective distribution channel for providing support and building community. I think there are seven simple methods that can help build out your company's twitter presence effectively.

Create Great Content

Make sure that you're saying things or creating content that's worth reading. This should be targeted to your core audience and be interesting to read. Content can be links to your blog, interesting things you're reading, or conversation starters directed to Friend Feed. Whatever you do: Please don't constantly tweet about your service's purpose - it's not interesting or engaging.

Keep Tweets Shorter (< 100 Chars)

You want your Tweets to be short enough to be re-tweeted effectively. This means that you really have 100 characters to engage your audience so they can easily add their own thoughts and re-tweet (RT) your content. Doing this is key to increasing reach and increasing the re-production rate outside of your current follower base.

Provide A Way to Engage

I recommend directing the conversation to services that are better at managing threads (e.g., FriendFeed, A blog with Disqus). Providing this link allows people to engage with the content in a meaningful way. A few people like to comment and even more like to read the comments. This creates an even better destination for your users.

Time Your Tweets

Don't turn your stream into serious noise. Limit the number of tweets per day (Between 5 and 10). Seriously. Too many tweets will get you unfollowed fast. People want a clear signal from your company not one that they have to sort through for an hour. This also means that you should time content based tweets throughout the day. Don't load up 5 tweets at once. If necessary, use a service like HootSuite to schedule your tweets.

Track Your Tweets

Always track your tweets. Use one of the many services that allow you to shorten your URL to something that's trackable. One caveat: URL shortening services are becoming the next phishers paradise. I recommend that you use one of the well-known ones: TinyUrl or Bit.ly. (Shorter is better, that's why I think Twitter went with Bit.ly recently.) Monitor the success rates (clicks) of your content. Optimize your content and conversation for how people respond. Another service for follow / unfollow timing based on Tweets is: Socialtoo. That way you can see what tweets are effective (more followers) and ineffective (more unfollowers).

Curiosity Element

People click on things they're curious about or want to know more about. It's important to include one of these elements in your tweets that captures one of these two elements. Doing this will increase your click-through and engagement with your audience.

Personality

Twitter is more than a content distribution system. You should also converse with your followers, be helpful and provide fantastic, instant support. To this end, I recommend that you personify your Twitter account with a real person. This helps increase the connection between your company (as a person) and your audience (people too). After all, as I believe Tony Hsieh said, people connect with people. So, be personable.

Top 10 Landing Page Form Tips

Making great landing pages for lead generation is a science. It requires a lot experimentation using A/B or Multivariate testing. I'm always amazed how little things like changing the text, background or button color will increase conversion rates. But, in order to make lead generation successful, it's important to play with these things. Think of it as a game. There has been a lot on the web written about email marketing and landing pages, effective forms, color design, etc. I won't try to replicate all that goodness. Instead, I've assembled a list of things I think about when creating high conversion landing pages:

(1) Single Page - Everything your user needs to see, right there, above the scroll. Never make it a step form for lead gen. Registration is a different beast.

(2) Forms - Make sure the eye moves quickly downward (i.e., don't put form text to the left of the box). There's been a lot of study written about forms. Just make sure the user's eyes don't have to work to figure out what you're asking them for. And, don't add a lot of weird error checking here. Get the email address correct and move on.

(3) Fields - Limit the number of fields shown / required. Only collect what you NEED, not what all your sales people hope, wish and dream of at night. More fields or the expectation of more fields typically increase drop-off / abandonment.

(4) Big Buttons - Make the button simple, bright and easy to find (this plays into email marketing as well). Make the button action reflect the call-to-action. Don't just make it "Submit." Make it "Download Free Copy" or "Get Free White Paper."

(5) Simple Messages - Do not include too many details that would detract the user from proceeding (e.g., don't get caught up in limitations, complex value props, long sentences, long quotes, etc.). Big clear value props shown to the user in first person form.

(6) Add Aspirational Context - Show the user other organizations / logos / quotes that he or she wants to emulate. Make the user think: "Wow, Wal*Mart has world-class, cost-effective IT. I could use this to cut the costs my boss has requested."

(7) Privacy - Couch any privacy concerns that a user may have. Give them the warm fuzzy - "We don't spam. We respect your privacy. We won't sell you out to our quasi-partners." Oh, by the way, only do this if it's true.

(8) Sharable - Add a widget from SocialTwist or Gigya that allows them to tell other friends. This should be shown on the download page as well (or if you initiate the download directly from this page - that's good too)

(9) Show Pictures - Pictures that load fast are required (typically, changing pictures or some other motion will keep interest to finish the form). Some of this is demographic dependent. Younger people like changes, older people seem to like static. A lot depends on the digital native / immigrant notion of your target.

(10) Test, Test, Test - Make sure that you test the hell out of your campaigns. Learn from each iteration to tune the form for you audience. Every .01% counts in making your lead or customer acquisition cost go down.

Do what you say you will do

Today, I ran into a situation where a leader is failing to deliver on a commitment. Sadly, this happens often with managers or business leaders. Ultimately, it erodes the person's ability to lead effectively. This reminded me of one of the mantra's that stuck with me from The Leadership Challenge by Posner and Kouzen.

They have a mantra in the book called: "Do what you say you will do," or DWYSYWD (note: it sounds more like a development tool than a leadership tool, but I digress).

It's simple: Don't commit or say something that you're not will or cannot do. This sounds simple, but is often difficult for leaders when they're on the spot. Often less capable leaders will make commitments he or she is not able to follow through on. Living by this mantra requires reflection and thought before speaking. The weight of one's communication will make or break the respect his or her team has for them.

So, before you commit or speak, make certain you can deliver. Otherwise, articulate your intent but not the definitive outcome or promise (implicitly or explicitly). This, of course, is only one aspect of leadership, but I think it's one that leaders often spend less time thinking about.

Instrument. Measure. Test. Optimize.

Web 2.0 is a big experiment. The corpus of knowledge is under developed. Some argue that the current web world is really only 3 or 4 years old. I agree that web as a platform is young and still acts like a child finding its way in the world. Best practices are still being developed. There is certainly no clear clear path to success. Granted, there are guidelines and companies that one can emulate. But, every business is a little different. Success first requires failing quickly. I've spent a lot of time talking to and listening to people who've created successful web companies and services and it really comes down to four activities beyond ideation and initial product: (1) Instrumenting User Interactions / Paths; (2) Measuring Drop-offs / Abandonment; (3) A/B & Multivariate Testing; (4) Optimize and Change Rapidly.

Unlike other industries, the data from web businesses is abundant and easy to collect. Tracking user interaction and engagement is trivial. Performing tests of new appflows is cheap and easy. One must be constantly hypothesizing and testing variations. Rigor in this testing is what drives success.

Instrumentation

Since I come for the systems and application management world, I understand the complexity and need for deep and meaningful user interaction instrumentation. There are a lot of companies out that that help make this a development after-thought (e.g., Omniture, Tealeaf). Yet, much of the cost can be eliminated if you make instrumentation an initial development requirement. The key is ensuring that elements are properly tagged and easily traced and groked from logs. If instrumentation is a core building block from the beginning, the rest of this process becomes easier.

Measure

Identifying what's important depends on the web application. However, the most important metric is funnel drop-off based on the goal of the business. If you're an retailer, then shopping cart abandonment is paramount. If you're a social site, then sign-up is paramount. If you're a freemium based service company, then conversion rate (or drop-off) is paramount. If you're a word-of-mouth site, then invite quantity and funnel is paramount. If you're a lead generation site, then form complete is paramount. Identifying the most critical flows to one's success, adoption and user engagement will drive these measurements. Not only is it important to measure drop-off rates, but it's also important to begin to understand why certain user behavior is occurring. Each of these measurements should become a daily dashboard metric that product teams live by. Subsequent experiments should be developed and continually tuned.

Test

Experimentation is what drives success. Initial educated guesses are typically wrong. And, web product teams that are not driven by statistics will likely fail to hit key milestones. Statistically driven experiments is what makes for better web applications and user adoption. One must define experiments in A/B and Multivariate terms. Doing so ensures that one can eliminate user drags and optimize user lifts. Simple things such as: color, button placement, number of fields, placement of field text, select alls can drastically change adoption. There is a lot of literature on these topics but the key to success is making this a priority often over new feature development. Remember, if you don't have the users or a clear adoption or conversion rate, then why does the next new feature matter? So, develop intelligent and data guided experiments that guide statistically significant change.

Optimize

Iteration is the name of the game. Taking small steps often is key. The faster the release process the better on the Web. Releases don't have to be large, but they should be focused with clear goals for measurement and testing. Having scientific discipline around this process from a product and development standpoint will help make one's web products successful.

At the end of the day, there are 150 Million websites all vying for users. If you're not constantly innovating and testing, the site is likely to fail.

Web BETAs : Top 5 Communication Requirements

As someone who signs up for almost every BETA under the sun, I've seen five best practices emerge that seem to keep me engaged throughout the process. Since I'm a Marketing person, I look at this from a product manager / outbound communication perspective. There are a LOT of things that should be done internally with engineering and support during this process to make your BETA cycles effective.

Internet BETAs have reshaped what BETA means. Traditionally, by definition, BETA meant a closed, invite only perfecting cycle before release. Well, in the web world of today, we now have private BETAs, public BETAs, and always BETAs. Always BETAs (and likely public BETAs) should really simply be released services, not some CYA low-quality service excuse (e.g., Gmail). Anyway, if you're in either of the first two types, you should add these five things to you BETA project plan.

Email Registered / Interested Users

The communication coordination for most BETA launches is horrible. The product / marketing guy typically misses the email announcement boat by a couple days. I typically hear about a BETA launch via a blog a day or more before I get an announcement email. This doesn't reflect well on your company or your service. Make sure that you send an announcement email before or shortly after you've pitched the bloggers. If you're doing waves of invites, you should still let anyone who was originally interested in your service that you've started sending out BETA invites.

Re-Introduce the User to Your Service

A lot of BETA invites / announcements I receive really don't tell me much about what it is the service does. And, that doesn't mean you say, 'it's a social network that works where you are.' Remind me what I can actually do with your service (i.e., what problem is this going to help me solve again?). This re-introduction should occur in both the email and as a splash page / getting started section on the website. Otherwise, most users can't remember what you do, why they signed up 4 months ago, and why they should consider giving you feedback. (Remember: the goal of BETA is go get great user feedback).

Invite Active Users to BETA User Groups

A lot of the early adopters (digirati-types) are not representative of your target market. That's the market that will actually make you money in the medium-run. Yes, they're the gatekeepers so you have to make them happy (e.g., make firefox and safari usable), but try not to spend valuable cycles on integrating with the latest / newly launched service. You should try to identify semi to active users who you believe represent your target market. Invite those users into a special BETA group. Then incentivize them to work with you to provide detailed feedback or to act as a sounding board.

Make Social Features Available for Viral / Word of Mouth

Everything on the web is becoming social. Your BETA process should account for this and make certain that your social features are implemented in a way that allows users to invite friends. Invites should be limited (5 per person) to manage growth and scaling problems. But, this will ensure that users get the full use of the site with the people they interact with most - friends. This will greatly increase engagement and frequency, which ultimately improves BETA feedback.

Open Feedback Channels

Make sure that you have a well-defined user feedback / support process prior to public or always betas. For private betas, you must make sure that the entry point for this process is well-defined and ideally a sorting and bucketing function exists so you can get prioritized feedback to engineering for your next iteration. The more open and transparent the feedback process is the better your users will feel about providing good feedback. No one likes sending well thought-out information into the blackhole of a company.

Innovator's Dilemma

Knowing when to launch a new product is tough. Is it ready? Does it work? Will they buy it? You must find the right point where the product is good-enough. If it's too good, you've lost timely revenue because you could have launched it earlier.

What is good-enough depends on your customer. In the web world, the customer is pretty forgiving. (Com'on Gmail is still in Beta; Twitter is up only slightly more than down; EC2 and other SaaS are at 2 or 3 9's.) And, it's better to get it out there to get feedback in the wild-wild-web. You can throw something out there that works, isn't scalable and consider it beta. (Not sure why people have created this notion of Private Beta - Beta should have an invite / approval only process, which means it's inherently private). Seems easy.

Now, enter the less-forgiving world of enterprise software. Granted, if you have a stable of devoted customers who love your products the bar of good-enough is lower (Think Microsoft's Technology Adoption Program - TAP). But, in the 'new product' or 'young company' world, good-enough is more difficult to assess. Finding companies or people to test your product in a Beta phase is more difficult. You have no track record. Thus, you want to make certain your product is to a level where it's not too burdensome to the beta tester.

But, most of the time people release a product after it is better than good-enough; but, that's much easier to assess in hindsight than looking out at the market. Remember: You have about 5-6 months after launching a new product before you see real sales traction; especially if you're a young company. So, if you screw up the GA release, you can always issue a patch a month or two after the release to clean-up any rush-to-market problems. Also, the following assumes that you're satisfied with the requirements that actually made it into the release plan.

Here is a list of decision criteria that I use to assess the good-enough factor of when to GA a product:

Installs :: Unless you have a stable of professional service people at your calling, then you need to install a product that the typical competent, relationship-savvy sales engineer is capable of installing. Sure, if the product is brand new, you have engineering and QA to help out. But, if not, it's typically a good idea to make sure that the product installs without too much heavy lifting.

No Blockers :: You must assess whether or not certain features work. If they don't work, you probably don't want to ship the product. If they work pretty well, and seem to be becoming less buggy as you approach your ship date they're good to go. But, deciding what is a blocker is an important discussion with your engineering, QA and support teams as we PM. Everyone has a different perspective on what's a blocker and there's always some grey area on what's a critical bug and a blocker. But, at the end of the day if you think your prospects or customers will have a serious problem as a result of the bug - it's a blocker!

Elicits Feature Requests :: You want a product that installs, is usable and elicits feature requests. If during your beta cycle you don't seem to be at this tipping point, then you should probably hold on to until it is better. But, if the product is eliciting feedback and feature requests, then you're likely to be at a release ship point.

Bug Fixes Beating Bug Filings :: Product releases reach a turning point where the number of bugs filled per day is beaten by the number of bugs fixed. This is an early indicator that the product is starting to reach a release point. However, one must watch the aggregate number of bug filings to fixes to ensure that the product has reached a quality point to release.

Does No Evil :: Make sure that your product does no evil. The more intrusive the product the more careful you have to be. For example, if your product requires super-admin privileges to function then it better be very well tested. And, more importantly, if your product changes something, then it better be rock solid and not do anything that will create work or headaches for your customer. Understanding the 'change' and 'impact' factor of your product is key. You typically only get one shot with post-beta customer, so you want to ensure that you get a good Beta group to mitigate post-GA hiccups.

Top 5 Product Management Mistakes

Product managers (PM) are stuck in the middle of a lot of activities. This makes PMs prone to making mistakes because ensuring priorities are always maintained is challenging. Great PMs are focused context switchers with strong system awareness (i.e., able to see the whole picture, even when deep in one task). There is a constant pull on time and it's easy to get caught up in one task and ignore others. In my opinion, the top 5 mistakes most PMs (myself included occasionally) make are failing to do the following:

1. 360º Communication :: PMs must act as the communication hub for the product. A clearinghouse of information and orchestration. The level of orchestration required depends on the size and process maturity of the organization. And, if you're a one product company, you're often the detailed communication hub for the company. This means that PMs must listen carefully as well as speak constantly to most stakeholders internally and externally. Evangelizing strategy and vision are as critical as ensuring that disparate groups (often silo'd) know what each group is doing; PMs often help facilitate conversations that help products function well. Moreover, communicating upward and outward is also important for ensuring that you're successful as a good PM. At the end of the day, you may be sick of the story, but you have to keep telling it - consistently! And, remember, know you audience and craft your message carefully. It matters.



2. Identifying Market Changes :: PMs can often get bogged down in the day-to-day tactics and lose visibility of market dynamics and strategy. Coming up for air and doing a quarterly or bi-yearly market analysis is not good. PMs must spend time daily consuming competitor and markets news and information. Setting up Google Alerts (and actually reading them) is important to consuming this information. Constantly consuming this information as well as following a "One Customer A Day" policy allows you as a PM to stay in touch with the market's changing needs. This helps drive strategy. And, it'll make writing MRD's a lot easier down the road.



3. Managing Expectations :: Managing expectations is one of the most challenging aspects of product management. Many PMs fail to do this because it's difficult and takes time to figure out how your organization works. In developing new products, there are a lot of unknown variables, e.g.: research & development time, testing time, feature changes, market shifts. As a result, properly estimating many of these is difficult as there are a lot of moving and dependent parts as well as stakeholders required to innovate. Not to mention, inherent in experimentation is the risk of the unknown. The more experimental the product, the more likely it will slip. So, good PMs are able to set expectations low and realign expectations periodically. But, great PMs can put pressure in the right places within the organization when required. Managing expectations with customers is likely the most critical aspect. Don't commit to sales people or customers that a feature will be delivered in a release until it's in Beta. And, if you have to commit, make certain that you make the feature a very high release blocker. Build trust. Finally, build levers within your launch plan that force stakeholders to come together and delivery when required.



4. Tracking Development :: Certain PMs write requirements and trust they will be delivered. Well, what's written is rarely understood exactly as intended and can be easily interpreted as being far more complex or less complex than required by the market or your customer. Sure, well written requirements mitigate this, but keeping track and setting-up tollgates within the organization to ensure words manifest into features one intended is critical. So, make sure as a PM you go as an observer to design reviews and work with QA to create -- or at least certainly review -- test plans that ensure requirement traceability occurs. Also, as features change or get dropped, make certain that that dependent product chucks are updated accordingly. And, of course, the more guidance and questions you answer early in the development process the more successful your product will be at meeting the market's needs.



5. Ruthless Prioritization :: Many PMs get lost in tasks. There is a lot to do. Failing to ruthlessly prioritize your workload means that things get missed and you end up working on the wrong things at the wrong time. This include product and feature prioritize as well. It's critical to go for what's important, not what's nice. So, make sure you prioritize ruthlessly in order to maintain product progress and work-life balance. And, if you are changing your priorities, make certain you do the other 4 things on this list accordingly. Otherwise, you've just trapped yourself into one of the common mistake areas.

All of these are premised inherently on self-awareness and reflection. It's good to take time-out while at the gym or while doing something menial to reflect on your work, stakeholder relationships and how everything is being perceived.

Updated: Added the product feature to the prioritization mistake.