100 Tips for Better Beta Tests

100 Tips for
Better Beta Tests
An eBook by Centercode (and friends)

v1.1

CENTERCODE | 800-705-6540 | INFO@CENTERCODE.COM

100 TIPS FOR BETTER BETA TESTS | 2

Introduction
Managing a beta test isn’t easy. When you’re brand new to the job, you quickly realize that there aren’t a lot of standards or resources to guide you. This
means that your approach will likely be forged through trial and error. And, even after running many successful tests, there will still be smart ideas
you haven’t adopted because the pressures of product release don’t afford you much time to experiment.
We’ve developed this eBook to counter those challenges. Ultimately, it’s part of a broader effort to distill the fundamentals of highly effective beta test
management into freely available resources. But with an effort that big, you have to start with the basics.
We began by collecting some of our own best practices from ten years of providing beta test management solutions to companies of all sizes, developing
technology of all types. Then we realized how helpful it would be to add additional perspectives. Our clients approach beta tests, UATs, and CATs in ways
that are often as diverse as the products they’re creating, and they all have strengths that others could learn from. We were lucky enough to get input
from many of our clients including Adobe, Autodesk, Avid, TiVo, Symantec and UPS, all fantastic companies with great people leading their test efforts.
Our hope is that this compilation of tips and tricks will help you become a more knowledgeable, productive, and confident beta test manager. Whether
this is your first test or fiftieth, we think we’ve covered enough material that you’ll find something of value here. In case we missed something, we’d
be happy to take more contributions (via comments on our blog or tips@centercode.com) for the second edition.



Basic Terms
Before we get started, we’d like to clarify a few basic terms used throughout this eBook.

BETA TEST

APPLICANTS OR CANDIDATES

When we say beta tests, we’re using the term generally. We just mean

We use these terms interchangeably to indicate users who have

a test where you introduce your product to a group of people who are

demonstrated an interest in participating in your beta test or program,

similar to your target customers and will use your product in their

but have not yet been selected as beta testers. Similarly, we refer to

real-world environments. These tests go by many names, including

alternates as those customers who met the qualifications, but were not

beta testing, field trials, pre-release, customer validation (CV), customer

initially included in the primary beta test team.

acceptance testing (CAT), and user acceptance testing (UAT). To many
companies, these tests have subtle but important differences. However,
they all share the basic idea we defined above. So, for simplicity’s sake,

BETA TESTER OR PARTICIPANT

we’re using “beta” to refer to all of them.

We also use these terms interchangeably, referring to users (generally
representing your actual target customers) who were selected to participate

BETA PROGRAM

in your beta tests. We refer to the collection of all beta testers within a
project as the tester team.

When we use the term beta program, we’re referring to the collection of
beta tests managed at one company, whether that means beta projects
for multiple products or past and future revisions of a single product.

BETA SUPPORT TEAM

Generally, when referring to a beta program, we’re discussing topics

This term refers to your internal support team for the beta, which may

with a scope that’s broader than a single test.

be no more than a single product manager or quality engineer, or might
include an entire team of stakeholders from a multitude of organizations.
Generally, they’re the people reviewing and responding to feedback and
communicating actively with participants throughout the test.



Tip Outline
1.

Planning Your Beta Test 1-12

8.

Maintaining Participation 65-75

2.

Recruiting Beta Candidates 13-24

9.

Handling Feedback 76-81

3.

Selecting Beta Testers 25-32

10. Dealing With Scheduling 82-85

4.

Handling NDAs and Agreements 35-38

11. Incorporating Other Teams 86-90

5.

Kicking Off Your Beta 39-42

12. Closing the Beta 91-94

6.

Assigning Tester Activities 43-51

13. Rewarding Your Testers 95-100

7.

Communicating With Testers 52-64



Planning Your Beta Test
1. START WITH A PROJECT PLAN

3. BALANCE YOUR CORE PARAMETERS

Like any well-run project, having a solid plan before you start is key. A

There are three core “moving parts” in every beta test: (1) the size of

good beta project plan includes (1) the objectives of the test; (2) target

your beta tester team, (2) the duration of your beta test, and (3) the

market details (i.e., beta candidate criteria) including their demographic,

set of specific goals that you’re trying to achieve. It’s useful to think

technical, experience, and geographic requirements; (3) beta tester

of these resources in equilibrium, where an adjustment to one has a

participation methodology and expectations; (4) the test schedule

countervailing effect on the others. Use this to your advantage in planning

(including planned build releases, time frames, etc.); (5) the intended

the most effective test. For example, increasing your test duration will

size of the tester team (broken down by market segment); and (6) a list

allow you to accomplish more goals. If your schedule gets cut, you can

of stakeholders and their responsibilities. Other details are great, but

often compensate by adding more testers and still achieve your goals.

these are all essential.

Factor these three parameters into your planning, but also keep them
in mind when unexpected events require you to make adjustments

2. SET REALISTIC GOALS
There may be many goals you want to accomplish during your beta, like
stressing certain features or testing different teams and resources under

throughout your beta.

4. EXPECT LAUNCH LAG

live customer action. However, you can only move so many mountains

Your testers lead busy lives, and they often won’t be ready to start testing

during a single project. If you think of each beta goal as a mini project

immediately upon receiving the product. For software, they may wait a

that requires scarce resources like time and the focus of your tester

day or two to install. If it’s a hardware product, they may not be home to

team, you’ll begin to understand why it’s important to space things

receive the shipment upon delivery. There are a variety of reasons that

out. Generally, we recommend specifying one named goal per week,

launch lag happens, but the point is that it does happen. To compensate,

in addition to basic test functions like validating quality and collecting

it’s useful to add a week or so to your plan.

general product feedback. If you attempt to accomplish several major
goals in tandem, you risk making little progress with any of them.



5. SIZE YOUR TESTER TEAM BASED ON
YOUR TARGET MARKETS

7. GET BUY IN

Most beta tests introduce a product to numerous target markets (or

collect almost always affects several different people and/or teams. Thus,

market segments), typically based on attributes such as region, gender,

your plan should also contain information about key stakeholders so

income, and technical knowledge or requirements. It’s important to keep

that they’re aware of their responsibilities throughout the project, key

in mind that the number of market segments you need to reach should

milestones, and general process descriptions. This generally includes

directly increase the size of your tester team. You don’t want to work in

product management, QA, and support at the very least, but may also

the other direction and select a number of testers to recruit, then hope

include product marketing, sales, and members of the executive team.

Beta tests are rarely managed by one person alone, and the data you

you’ve adequately covered your target market. If the composition of
your tester team doesn’t bear an accurate relationship to desired market
segments, it’s difficult to weigh the relevance or importance of survey
results (i.e., they become anecdotal).

8. DON’T FORGET ABOUT RAMP-UP TIME
If you’re starting a beta program from scratch, recruiting a great tester
team can easily take two weeks or more, depending on your target

6. BASE PROJECT LENGTH ON GOALS

market requirements and the size of your test. If you’re starting a beta
project with either an existing (hopefully interested) customer list or

As a baseline, your beta test should be no shorter than two weeks (3-4 is

an established beta community, ramp-up can be reduced to only a few

generally optimal). Beyond that, the length of your beta should be tied to

days. Either way, it’s important to include this period in your plan. The

your project goals (and to some degree the complexity of your product).

last thing you want is to sacrifice planned testing time to make your

We’ve already discussed that under most circumstances, you should be

product release window because unanticipated recruitment delays

pursuing one specific goal per week. Thus, if you have four primary goals

consumed 25% of your beta schedule.

that you want to accomplish, your beta test should be at least four weeks
long. If you need to achieve more goals in a shorter period, consider
increasing the size of your tester team and splitting your tester team into
focused groups. That way, you’re maintaining equilibrium among your
core parameters, and your tester team’s attention isn’t being diluted by
trying to address several simultaneous goals.



9. BUILD IN MULTIPLE PHASES

11. PLAN FOR CHANGE

Splitting a beta test into phases (e.g., Beta 1, Beta 2) offers a number of

While your plan is a great starting point, beta tests quite often change

advantages. One is the ability to slowly introduce a larger tester team,

course rapidly and unexpectedly. You may run into show-stopping

which allows you to reduce the impact of early bugs, ultimately burning

software issues that require additional phases or participants. Another

out fewer testers. Another benefit, specific to hardware tests, is the ability

group may run into an issue that delays a build by a week or more.

to cover more of your target market (and their unique environments)

Early feedback may change the primary goals. Be prepared to adjust

with fewer expensive, pre-production units by redistributing hardware

as necessary, communicate changes clearly to all involved (especially

between phases. Note that if you require time between phases (a few

changes regarding the beta test schedule or goals), and always update

days or more), communicate this clearly with your beta testers to ensure

your plan accordingly.

they remain aware and engaged. It’s best to keep this downtime to an
absolute minimum when possible.

10. PLAN FOR IDLE PARTICIPANTS

12. PUT YOURSELF IN YOUR TESTERS’ SHOES
“Plan. Think things through, pretending you are the beta tester. What
do you need to help you test and report information back? Then ‘beta’

It’s extremely uncommon for every beta tester to meet the goals you’ll

your test with a few people to make sure it works as you expect and

set for them. Sometimes participants are simply unmotivated (which

that you are getting back the information you need to make your testing

other tips in this guide address directly), but many times other personal

successful and worthwhile.” -Gayle Musker, UPS

or business responsibilities take precedence. It’s crucial to factor this into
the recruitment section of your beta plan. If you’ve never managed a
beta test before (therefore having minimal recruitment and participation
management experience), you should plan to include at least two to
three times the participants you consider necessary to meet your goals.



Recruiting Beta Candidates
13. BUILD A POOL OF TARGETED CANDIDATES

15. PROMOTE EXCLUSIVITY

While public betas that anyone can join can be a useful marketing activity,

It’s a good idea to let testers know that not everyone will be selected.

traditional beta tests are better suited for a more limited audience.

For one thing, it reduces the feelings of disappointment (and sometimes

Rather than simply letting anyone in the test, it’s best to build a large

Internet rage) among those who were not selected. However, it also

pool of applicants from which you’ll select those who both best meet your

emphasizes the importance of participation right from the very first

criteria and exhibit the signs of great testers. Remember, just because

interaction with your beta project. If applicants know that they’re signing

a beta is public doesn’t mean you’ll get more feedback, but it does tend

up to become one of a limited number of participants, they’ll understand

to make the feedback you get harder to analyze because you’ll typically

that there’s a higher degree of responsibility with that application.

know much less about the testers submitting it.

16. START WITH AN APPLICATION
14. RECRUIT MUCH MORE THAN YOU NEED

It’s almost always beneficial to present interested candidates with an

Not everyone who applies for your test will be a great candidate. Some

application survey containing 5-10 questions that they’re required to

won’t meet the basic technical requirements; others won’t match your

complete. The goal of every question should be to help you identify the

target market; and some will demonstrate a lack of skill in clearly

best testers for your specific product. Generally speaking, what you’re

communicating (thus making your life much more difficult down

intending to do is identify candidates who: match your target market;

the road). It’s generally best to recruit at last three to five times more

are responsive and effective communicators; pay attention to detail; and

candidates than you actually plan to select, allowing you to choose only

are genuinely excited about your company, product, or similar products.

the most qualified beta testers. If you can get more, that’s even better
(we generally aim for ten times as many).



17. RECRUIT WITH OPEN-ENDED QUESTIONS

20. SET PARTICIPATION EXPECTATIONS

One great way to increase the value of your application survey is to

It’s crucial to always let your testers know exactly what’s expected

include a couple of open-ended questions. Asking something like, “Why

of them in a beta test, from reporting bugs, to completing surveys, to

do you want to beta test this product?” will act as a mini-interview

participating in forum discussions with other participants. We find it

process, allowing the tester to demonstrate the type of effort they’re

best to first set expectations during recruitment, again after selection

likely to put into feedback during the beta itself. These short answers can

(at the test kickoff), and once every couple of weeks throughout the test.

be invaluable in narrowing your pool to the most effective candidates.

21. IDENTIFY ANY COSTS
18. KEEP YOUR RECRUITMENT VAGUE

Beta testing is voluntary, and as a general rule, should incur no costs to

You want people fresh and unbiased when they receive the product. Your

participants. If there is a cost of ANY KIND associated with participating

recruitment messaging should entice them to sign up while hiding the

in the test (for example buying paper/ink for a printer being tested or

details of the actual product. That way, they’ll be excited and interested

data rates on a mobile phone), go out of your way to cover the costs

in exploring the product once it arrives. Furthermore, it provides you

yourself. If for some reason this is impossible, make it very clear to

flexibility and eliminates issues with tester expectations.

every tester before they sign up.

19. STATE CLEAR REQUIREMENTS

22. SET MANAGEABLE TIME COMMITMENTS

If your product (or beta phase) has specific requirements (hardware,

When a beta program has participation problems, it’s easy to view testing

software, demographic, geographic, or experience/knowledge), make

time in extremes. You grow accustomed to seeing many participants

certain that they’re clearly communicated to all test candidates. If

give little to no time, while others are heroes who treat the beta like

essential, take extra steps (such as additional surveys or even personal

a full-time job. What you want to avoid is letting that mind-set carry

phone calls) to verify that candidates meet the requirements. Recruiting

over into your beta plan by building in significant time commitments

testers who literally cannot participate (no matter how bad they want

but expecting only a few to keep them. Instead, establish a realistic

to) is a huge headache for everyone.

commitment of time and effort from each beta tester but expect that
they will satisfy it. It also doesn’t hurt to plan special rewards for those
that do go the extra 10 miles.


23. USE SOCIAL NETWORKS
Social networks are an awesome way to find willing participants.
Twitter, Facebook, and LinkedIn all provide you with a free and simple
mechanism to track down great candidates. If your company has active
social media marketers, start a dialog and enlist their help. If you have
to go this route alone, tools like TweetDeck, CoTweet, and HootSuite
allow you to monitor conversations relevant to your product across
social networks.

24. LOOK OUT FOR VIP TESTERS
“During the prospecting process, it will become evident that some
customers are more likely to provide good feedback than others. A subset
of the beta participants will warrant extra attention and support during
the program as they are more likely than others to help you reach your
targets. Identify these premier companies early. Make all internal parties
in your company who are involved in the beta program aware that these
special firms will be monitored closely, and if needed, given additional
help and support to ensure they are successful in the beta. Assign named
technical resources to these sites to better ensure their beta program
success. The product management team should work especially close
with these customers, visiting them or engaging in additional contact
to help resolve issues that arise.” -Jeff Crawford, Adobe




Selecting Beta Testers
25. BE SELECTIVE (SIMILAR TO #13)

27. CONSIDER APPLICANT ENVIRONMENTS

Don’t let just anyone into your beta. Public beta tests sound like a

Most technology relies on adjacent or parent hardware to function

seductive idea because you don’t have to spend time reviewing and

properly, like the PC that runs your software or the mobile phone that

selecting testers, and an open beta test has to result in more feedback,

runs your app. These environments beyond your control can cause many

right? Except, it doesn’t really work that way. Public betas are hard to

issues and bugs for your product, and are one of best reasons to run a

manage, and you end up with a larger beta that gives you less feedback.

beta. So, when it comes to selecting your participants, you want to know

With a public beta, people generally join out of curiosity, which is a

more about them than just personal characteristics. It’s also important to

much different mind-set than you get in closed beta tests. In addition,

know about the environments (we call them “test platforms”) that they’ll

it offers a much wider group of potential customers an early look at an

be using during the test. Ideally, you want to be testing your product in

unfinished (and most likely quite buggy) product that can present a very

dirty environments that resemble what it will face when customers start

negative impression, ultimately hurting sales.

taking it home (or to work). Then you can analyze the feedback you get
in light of that test platform. This is one of the primary (and most well

26. LOOK FOR DETAIL AND CARE

understood) benefits of beta that you don’t get in the very controlled,
quality lab environments.

Select candidates who complete a test application with great detail, limited
spelling and grammatical errors, and a good argument for why they
should be in the beta. These people are likely to deliver similar results

28. SEEK OUT PASSION

during your project. On the other hand, PEOPLE WHO WRITE LIKE

If someone who meets your basic requirements is hounding you to get

THSI SHULD B AVOIDED (unless, of course, that’s your target market).

into a beta test, you should give that person a chance. Enthusiasm and
persistence are two excellent traits in beta testers, and that person will
likely be one of your more active participants.



29. AVOID FRIENDS AND FAMILY

31. MAKE SELECTIONS QUICKLY

The ideal beta participants are those who are unbiased and truly

The availability of your testers can change, so we recommend keeping

represent your target audience. This effectively eliminates friends

the window between recruitment and beta tester selection very short

and family, as well as most employees, from being great candidates.

(days, if possible). This will help ensure that those who volunteer are

Sometimes their loyalty compels them to hold back, while other times

still excited for the project and still capable of committing the time

they don’t feel obligated to participate like other testers. Either way,

necessary. Otherwise, you might find yourself reopening recruitment to

your best move is to rely on testers that you don’t already share a prior

find replacements for the testers you lost, further delaying your project.

personal relationship with.

32. DON’T TRUST EMAIL
30. AVOID SALES LEADS AS TESTERS

If you identify a highly qualified candidate, but they don’t respond to

It’s common for sales to ask a beta manager to let a lead into a beta

your invitation to test, it’s likely they missed your email. Often, a simple

test. There are at least two problems with this, though. First, sales leads

phone call will reveal that your message was missed or ended up in a

usually aren’t interested in testing your product, but rather they’re

spam folder. People generally don’t express interest in being in a beta

interested in evaluating your product. They’re looking at it relative to

test and then immediately turn around and change their minds.

their immediate needs, and you can’t rely on their participation because
the motivations are completely different. Second, beta testers are more
willing to forgive bugs and other quality problems than a customer who’s

33. ALWAYS KEEP A POOL OF ALTERNATES

focused on evaluation. Besides, the last thing you want is an angry call

There can be many reasons for losing a tester. Some realize they

from sales insisting that their lead didn’t buy the product because it was

don’t have the time to commit, while others have to be removed for

too buggy during the beta.

disciplinary reasons. Whatever the reason, it’s a good idea to have a
pool of alternates available. You don’t want to go through the hassle
of reopening recruitment in the middle of your beta test. And having
alternates also gives you the ability to add a few extra testers if it looks
like you won’t be hitting your participation goals.


34. NOTIFY SELECTED TESTERS ONLY
You may have the urge to let the applicants who weren’t selected know
that selection is complete. We’ve found that this is usually a bad idea.
More often than not, those who weren’t selected will react negatively,
feeling snubbed because they weren’t selected when they feel they’re
perfectly qualified. It’s generally a better idea to set expectations ahead
of the recruitment, with a message along the lines of “You will be notified
within one week if you’re selected to test.”




Handling NDAs and Agreements
35. ALWAYS USE AN NDA

37. SEND OUT NDA REMINDERS

This may not need mentioning, since companies seem to request a

Sometimes, especially in longer betas, it’s a good idea to send out a

non-disclosure agreement (NDA) before they’ll even tell you where

mid-test email about the terms of the NDA agreement. We find that a

their restrooms are. However, in the context of beta testing, NDAs are

good time is after the first few weeks or if you have a break in testing

critical. They offer you legal protection, but more to the point, executing

between phases. Basically, when enough time has passed for excited

an NDA (accompanied by a plain English explanation) reinforces the

testers to innocently forget that they can’t talk about the product, a

importance of beta secrecy to your tester team so hopefully you don’t

gentle reminder of the NDA terms and consequences of breaking them

have to enforce it. We’ve published a free kit that covers both NDAs

is all it takes to protect confidentiality and avoid an unpleasant situation.

and participation agreements for beta testing, including templates and
instructions, which is available at www.centercode.com/resources/nda.

36. USE A PARTICIPATION AGREEMENT

38. ENFORCE ZERO TOLERANCE
While it’s always on a new beta manager’s mind, beta NDA leaks are
actually quite rare. That said, when it comes to confidentiality, no

The role of participation agreements is a little less obvious than NDAs, but

violation is too small. If someone leaks anything about your project, you

still very important. They’ll cover things like participation expectations,

shouldn’t give an inch. Immediately remove the tester from the beta

receipt and return of the test product, limitations on what can be done

and notify the other testers of the action and penalty. While legal action

with the product, etc. While they’re binding legal documents, they’re

is generally not pursued in beta NDA leak cases (outside of industrial

generally not something you try to enforce in court. Their value is in

espionage, violations are generally the result of accidents or ignorance),

clearly communicating what’s expected from your testers and reminding

it is an option. Obviously this an uncomfortable series of events, but

testers who stray what they’ve committed to. The kit mentioned in

people need to recognize that you take the terms of the agreement very

the previous tip also contains a lot of information about participation

seriously. One crack in the dam is all it takes sometimes for the whole

agreements, including the key elements they should cover.

thing to come crumbling down.



Kicking Off Your Beta
39. START WITH BETA 101

41. CONFIRM CONTACT INFORMATION

It’s a mistake to assume everyone knows how to effectively participate

Before any test, and especially before shipping any physical product,

in a beta test, so start things off by sharing a set of simple guidelines that

remind all participants to update their contact information in whatever

explain what makes a great tester. You can also use this opportunity to

system you’re using. People move, change cell phone numbers, and have

introduce testers to the systems or tools you will use to manage their

email addresses they only use for signing up for things. You want to make

feedback (providing all relevant URLs), as well as provide tips for how

sure you have the most complete and current information available.

to write great bug reports, contribute positively in forums, etc.

Plus, shipping an unreleased product to the wrong house can be a source
of frustration and embarrassment (not to mention a waste of time).

40. REAFFIRM EXPECTATIONS
We mention expectations again because clearly communicated expectations

42. INCLUDE RETURN INSTRUCTIONS

are a big part of a successful beta test. Once the test is ready to start,

If you’re shipping physical products that you expect testers to return,

you should send out another message letting testers know what you

be certain to include clear instructions on how to return the product

want, when you want it, and how you expect them to accomplish it. This

and what you expect returned. Be sure to include everything necessary

allows you to establish a clear path to the incentives (i.e., if users are

to ship the item (prepaid, of course). Skipping this last step can greatly

doing what is asked, they get the reward). It also gives you a concrete

reduce the amount of product you eventually get back. It’s also helpful

reference point to leverage in future discussions with any testers who

if testers can ship back product in the same box it arrived in. It means

are not participating.

fewer shipping hassles for them, plus it gives them a distinct place to store
any instructions and return labels you provide at the onset of testing.



Assigning Tester Activities
43. USE THE RIGHT TOOL FOR THE JOB

45. USE SPECIFIC TASKS TO REGRESS FIXES

There are many types of feedback mechanisms. Beyond bug reports and

One area where a diverse and reliable tester team really shines is in

feature suggestions, other common activities in betas include the use

regression testing. If you’ve fixed some known bugs, verify that you’ve

of surveys, tasks, daily journals, discussion forums, wikis, etc. The key

solved the problem with a group (or, in some cases, all) of your testers.

is to think about your goals and select your tools based on what you’re

You can segment your team by test platforms that were known to exhibit

trying to accomplish. For example, if you know that documentation

the bug and assign tasks that follow the specific steps required to recreate

feedback is critical, that’s a great opportunity to enter your docs into a

the issue. Or, you can set your entire team after the problem just to make

wiki and crowdsource improvements.

sure it’s really gone. The added benefit is that testers will experience
the results of their efforts firsthand, leading to increased participation.

44. USE GENERAL BROAD TASKS TO
ENCOURAGE PARTICIPATION

46. ASSIGN OBJECTIVES, NOT STEPS

Some testers lack the initial drive to independently explore your product

“Goal-based test scripts are a much better way to get a thorough assessment

and report back their findings. We’ve found that giving people a set of

of your software’s usability. If you give them a task like ‘Turn on the

very basic, general tasks will help kick-start their use of the product,

Scheduler,’ you not only assess how the Scheduler works, but how easy

after which they’re more likely to do their own exploration. Note that

it is to find and use it.” -Tony Weiss, Symantec

these should not include tasks that will focus the tester on very specific
features or activities, but rather the product as a whole (i.e., download
the software; load the software; review the online help). In most cases,
while you may have to nurture participation in the beginning, testers
will be much more independent once they build some momentum.



47. GIVE REASONABLE DEADLINES

50. LIMIT YOUR SURVEY QUESTIONS

Remember, the people volunteering to test your product are (usually)

When testers are presented with a long survey, they may decide to

doing this in their free time. It’s important to respect that and give them

skip it just on looks alone. It’s not necessarily important whether the

a real opportunity to finish the activities (surveys, tasks, etc.) assigned

questions are simple or complex. Thus, we recommend limiting surveys

to them. We find that 2-3 days is reasonable for relatively simple tasks,

to between 5 and 15 questions. An even better approach is to only display

while a week is appropriate for more complex assignments. You can

the most pertinent questions by making them conditional on the answers

opt for shorter deadlines when necessary (and only sparingly), but

to earlier questions (a feature found in Centercode Connect). Not only

understand that completion rates will probably suffer for it.

does this give you the opportunity to craft a more detailed survey, but
it also makes the survey initially appear very short.

48. ISSUE SURVEYS OCCASIONALLY
Try to limit surveys to about one or two per week. They’re an incredibly

51. TIME ASSIGNMENTS STRATEGICALLY

useful tool in beta management, but there are consequences to assigning

If you’re planning a long beta, it can be a helpful to time your tasks

too many. Frequent assignments discourage your participants from

and surveys so they occur in the later stages of testing. Early on, many

exploring the product on their own, and if taken to the extreme, they

testers are likely to be excited and focused on exploring, not needing

will quickly frustrate and burn out your testers.

the additional push. However, as the test goes on, the extra direction
that assignments provide can often breathe new life into a waning beta.

49. AVOID TESTING FAILURES & LOST CAUSES
“Don’t ask testers to perform a test you know will fail. This just causes
unneeded frustration. Similarly, don’t ask your testers a question unless
you can act on the feedback. Testers want to make a difference (which
is why they volunteered to test), so if their feedback goes unheard, they
will be upset and will be less likely to volunteer for you in the future.”
-Amanda Dawson, TiVo

Tasks make it very clear what you are looking for from the tester, while
surveys tend to be an easy method of giving feedback (particularly when
they consist mostly of drop-downs, check boxes, and radio buttons).



Communicating With Testers
52. USE EMAIL SPARINGLY

54. GIVE PREFERENTIAL TREATMENT

You’ll want to ensure email communication is infrequent, important, and

“Pay extra attention to your super users. Some testers are like part-time

to the point. Consider limiting its use to messages that are either time

employees and will spend hours each day testing, helping others in the

-sensitive or very important. For other matters, stick with your beta test

forums, and evangelizing your product. Reach out to these testers with

tools (bug report comments, discussion forums, etc.) as your primary

a personal email, phone call, or extra gifts to make them feel special.

means of communication. Why? High email volume risks frustrating

It’s worth it in the end!“ -Geoff Griffin, TiVo

your testers and diluting the perceived importance of your messages.
Sending frequent emails also blurs the lines of your communication
protocol, encouraging testers to direct communication to you via email

55. ENFORCE COMMUNICATION PROTOCOL

instead of your beta test tools.

If you want to keep your beta test data organized, enforcing communication
protocol is critical. When participants have general test concerns, like a

53. OFFER POSITIVE REINFORCEMENT

change in their availability or questions about how to be a better tester,
it’s okay for that communication to happen by any means necessary.

Don’t forget to praise users for particularly good feedback or participation

Where you want to be strict is in how participants give you beta test

in the test. Comments like “excellent question” or “great thought” not only

feedback. If some people use your beta management tools, others send

make your users feel appreciated, they also reinforce that you’d like to

in random emails, and one person calls you to report bugs, things can

see more of that behavior. After all, it’s sometimes easy to overestimate

get out of control quickly.

how much guidance you’ve actually given your testers.



56. WRITE FOR THE APPROPRIATE AUDIENCE

58. CLICK SEND

It’s important to remember your audience demographics when

“Don’t be paralyzed by perfection with those emails you have to send to

communicating with participants. Take young adult gamers as an

lots of beta participants. You’ll have to send several of them to establish

example. They’ll tend to speak, engage, and operate a lot differently

good communication and follow up with your external beta testers.

than older business software testers. That doesn’t mean your tasks have

If you brood over the perfect language rather than just using your

to revolve around frags and pwning, but don’t expect them to provide

own casual, engaging tone (assuming you have one), then you’ll likely

detailed cost-benefit analyses for your product.

communicate too little, too late, or in a manner that puts off your audience.”
-James McKey, Symantec

57. TIME YOUR COMMUNICATIONS
“If you are running a global program, or even testing across country, be

59. STICK TO YOUR WORD

mindful of when you send the email based on time zones. Depending on

Keep your promises and watch your words. If you say something during

the target audience, you may find they have different habits as to when

a beta test, your testers will hold you to it. And this applies all sorts of

they check email and when they arrive at/depart from work. Don’t be

issues, including test schedules, product features, and (perhaps obviously)

afraid to ask these kinds of questions of your beta participants in exit

incentives. You might think testers won’t care if you have to go back on

surveys, and remember that you are asking them to help you free of

something you’ve said, but you’d be underestimating the personal stake

charge in most cases, so you need to flex to their schedule, not the other

they feel when it comes to testing your product.

way around. Scheduling outbound emails via macros based on time
zones can yield a high return rate if done right.” -James McKey, Symantec

60. KEEP THINGS UNDER WRAPS
Now that you know how important it is to watch what you say, what do
you do when testers ask hard questions? You still have the option to not
say anything. The point of a beta project is to gather tester feedback,
so as long as you aren’t ignoring the question or otherwise being rude,
you aren’t obligated to give full answers or any answer at all. The key
is to handle it in such a way that you don’t upset the tester and derail
future participation.



61. DON’T ANNOUNCE UPCOMING BUILDS

63. DON’T BURN OUT GOOD TESTERS

This one is a cardinal rule of beta testing. Never, ever tell participants

It’s natural to want to identify good testers and keep inviting them to

that you have a new build coming. They will stop testing and wait to

more tests. However, even active and engaged beta testers can suffer

see what the new build brings if they know a release is on its way. And

burn out. Keep a careful watch on their productivity if you use them

if you think about it from their perspective, that makes sense. Why

often, so you know when this might be happening. And if you want to

would they keep testing when anything they uncover could already be

keep your relationship going with the tester while still imposing a break,

fixed unbeknownst to them? You see it differently because you probably

consider using them as an alternate.

have the benefit of knowing what is being addressed in the new build.

64. CONSIDER HOME OR OFFICE VISITS
62. BE PROFESSIONAL, BUT BE HUMAN TOO

“Try arranging visits with beta customers in their test environment

Balance professionalism with personality when managing your tests.

(often, if you can). Invite product designers along for the trip, too. One

Not only will this help your participants feel like they are part of a team,

of the challenges of product development is verifying assumptions about

but it can also pay dividends when it comes to participation levels. It’s

how your users use your product. What better way is there to do that

hard for testers to relate to a stodgy and stilted beta manager, and if

than seeing for yourself?” -William Marshall, Avid

they can’t relate to you they’ll feel less compelled to help you.



Maintaining Participation
65. USE GOOD TOOLS

67. RELY ON OPEN-ENDED QUESTIONS

At the risk of self-promoting, beta tools matter. Giving beta testers access to

You can really encourage participation by relying on open-ended

good tools for sharing results helps ensure that you actually get feedback.

questions. When you ask simple “yes” or “no” questions, you’re closing

Tester-friendly and well-organized beta tools build momentum early in

off opportunities to learn more about what participants think of the

the test. Whereas, if your testers struggle to share data with you or have

product. They might have had great feedback if the question asked “why”

to juggle a bunch of different tools (and separate user logins), they’re

or “how” instead of “yes” or “no”. There are times when you might have

less likely to participate. There’s a reason why we see an average of 90%

to use simple, closed-ended questions — particularly toward the end

participation in our beta tests, and Centercode Connect is a big part of

of long beta tests, when participation is dragging — but these types of

that. Beyond reducing tester participation, a collection of disconnected

questions should be a fallback, not a staple.

tools also makes it much more difficult to monitor participation as a
whole, as well as work with the incoming feedback.

66. THROW OUT THE BAD APPLE

68. USE EXPEDITED SHIPPING
If you’re testing hardware, use overnight shipping if possible. The
added expense comes with benefits. First, it’ll help you ramp up the

One bad tester can spoil the whole bunch. If you have a person who is

test faster. Everyone experiences launch lag, but when your product

excessively abusive, negative, or offensive, it’s important that you pull them

spends less time in transit, you get shorter test phases. Second, it sends

from the test. Otherwise, you risk letting that person ruin the experience

a very positive message to your testers. By spending more on overnight

for your other testers. The more people are negatively affected by a bad

shipping, your testers will see that you’re just as eager as they are to

tester, the less likely they are to participate. So, effectively, you’re not

start testing the product.

only hurting your other testers but your product as well.



69. PRACTICE CREATIVE THANKFULNESS

72. AVOID OVER-RELEASING

We always encourage you to thank your testers early and often. They

It can be a fine line, but try to keep product updates to a minimum during

really respond when you show them you appreciate their feedback. By

your beta. On one hand, updating your product during beta does show

offering positive reinforcement throughout and incentives at the end of

responsiveness and gives you a chance to perform regression testing.

your beta, you’ll be well on your way to making your testers happy. If

On the other hand, too many updates can frustrate your testers and

you want to build an even more positive relationship, though, consider

discourage participation. Even simple software installations can get

adding something creative, unique, and commemorative into the mix.

tedious if you subject your testers to them regularly (hardware can be

Tester Team T-shirts, for example, show your appreciation and add an

much worse). And if you release updates frequently enough to become

additional sense of exclusivity to projects.

predictable, people will stop testing in much the same way as if you
were announcing upcoming builds.

70. BE A GOOD MODERATOR
This means being careful to watch for tangents and diversions. They’re a

73. DON’T DEMAND CONFORMITY

natural part of beta testing, especially when you run discussion forums

Beta testers are all unique. Sometimes it’s better to adjust the test to

for your testers. But if testers start to focus on one thing too long or

accommodate them than to force them comply with specific demands.

become mired in unproductive discussions, don’t be afraid to gently

For example, while user forums are a great way to keep users invested

steer them back to the charted course.

daily, some people just aren’t social enough to effectively utilize that
channel. In that case, asking for daily update journals are a better option

71. EMPHASIZE EXCLUSIVITY
One of the best tools at your disposal for building a sense of community is
the exclusivity of beta testing. Testers thrive on that exclusivity. It makes
them feel like they are themselves special as well as part of something
special. The result is a more dynamic community of beta testers that
are each more motivated to help shape your product.

that produces the same results.


74. ACT QUICKLY WITH IDLE TESTERS
The longer you let idle testers slide, the more it seems that you’re not
serious about participation requirements. Give your testers a gentle
nudge if you notice they haven’t been providing feedback regularly. A
simple email showing them what they’ve done and what you expect will
get most people engaged. If you let it linger, they will think you don’t
care and then it will be too late.

75. KEEP YOUR OPINIONS TO YOURSELF
Testers are easily swayed. If you express that you like or dislike anything
related to the product, they’ll demonstrate an aim to please. Suddenly,
your data will trend toward those opinions. If there’s a need to share
an opinion, be objective. Point out the good and bad of both sides or
ask questions that make testers think about the idea. The only time you
should use strong opinions is to encourage or discourage a discussion.




Handling Feedback
76. BE RESPONSIVE

78. USE BUG FREQUENCY FOR SEVERITY

One of the easiest ways to improve participation in your beta is to be

Redundant beta feedback may sound like a nuisance, but it’s actually an

responsive. Participants understandably get discouraged when they

excellent way to measure the severity of bugs. If a significant number

offer feedback but receive no response or even an indication it has been

of testers report the same problem, it immediately escalates its severity.

read. By being responsive, you indicate that their feedback is valued. It’s

Thus, it’s important to encourage your participants to report all problems

also an opportunity for you to get more information on reported issues.

they encounter. Doing triage on bug reports is a much better problem to
have than releasing a product with small, widespread bugs that slipped

77. BE GENUINE AND SPECIFIC
There’s a caveat to being more responsive: You shouldn’t fake it with
boilerplate responses to feedback. Giving the same canned response

through the cracks during beta.

79. SE A VARIETY OF FEEDBACK MECHANISMS
U

to all feedback is hardly better than no response at all. Users see it is

It’s good practice to offer your testers several ways to share feedback.

impersonal and it makes them feel like you don’t really care about their

Some people like discussing the product on forums where they can chat

input. Instead, make your response specific to the feedback at hand or

with others. Other people aren’t comfortable with the idea of bug reports,

the user’s participation in general (e.g., “Thanks for your hard work

but will provide amazingly detailed feedback in a daily testing journal.

recently. You’ve given us some very helpful bug reports and feature

Some thrive when given specific directions via tasks and surveys. By

ideas this week.”)

providing different opportunities to participate, your testers are more
likely to find feedback mechanisms that resonate with their individual
preferences and ongoing needs.


80. STAY OFF THE LOW ROAD
At some point, you’ll run across a participant who doesn’t like your product,
gets frustrated with a bug, or just has a negative attitude. This can be a
tough situation, but try not to suppress criticism or respond negatively
to it in any way. If you publicly censor testers, you risk alienating them
and limiting future feedback. And if you attempt to rebut what was said,
it looks like you’re not interested in honest and candid feedback. On the
other hand, if you feel like you could learn more about the problem by
asking questions, by all means do so. Just be careful of your mindset.

81. ENABLE TESTER CREATIVITY
You might think it’s best to discourage testers from using your product
in unintended ways, but there are benefits as well. When you allow
testers to use the product how they want, you’re tapping into a great
resource for future product and feature ideas. And since you can’t always
control how paying customers will use the product, it’s also a chance to
get insight into support issues that may arise from unforeseen use cases.




Dealing With Scheduling
82. BE REGIMENTED, BUT NOT SPECIFIC

84. INCENTIVIZE EXTENSIONS

It’s best to be vague about dates when talking to your testers. The nature

If you need to extend your test beyond the specified period, it’s important

of beta testing often requires you to shift your schedule to accommodate

to budget for additional incentives. Testers committed to specific

issues and changes. If you lay out everything in numbers of weeks, most

terms, and your good will with them can take a hit if you extend those

of what you’ve told testers will still be true if you have to adjust your

terms without sweetening the bargain. Just as important, the extended

schedule. If you give testers specific dates, one small hiccup cascades

commitment needs to be optional. It’s likely not their fault the test ran

into a dramatic change that affects every subsequent date.

long, and they shouldn’t be penalized if they can’t continue.

83. PRACTICE CONSISTENCY

85. MAKE THE MOST OF SLIPS

“Keep to a regular test schedule. Overall participation improves when

If you have already shipped a product or people are prepared to test

keeping to a consistent test schedule. I send out new testing instructions

but some last minute issue has caused the project to slip, you need to

on the same day each week and ask testers to complete surveys by the

change your focus fast. Examine a part of the product that is unrelated

same day each week. I’ll often see my numbers drop off when forced to

to the slip and ask testers to focus their energy on it. If the slip is severe,

diverge from our regular schedule.” -Geoff Griffin, TiVo

it may make sense to put the test on hold. However, only do that if you
believe that you can’t solve the issue within a day or two of test start. It’s
very difficult to regain test enthusiasm and momentum once a project
has been placed on hold.



Incorporating Other Teams
86. PROACTIVELY SHARE WITH OTHERS

88. BECOME A JACK OF ALL TRADES

We recommend finding time to share beta feedback with teams that

“Being a beta program manager requires some well-rounded skills.

might not be involved in the day-to-day efforts. For example, you can

You have to be good with customer facing engagement, proficient at

pass along feedback about documentation to your technical writers. You

marketing techniques, competent at managing a project, and able to

might also share product feedback and testimonials with your marketing

handle the details of following up with all customer issues/concerns

department. Reaching out like this helps bolster your organization’s

while compiling a detailed final report that can be ready for delivery

opinion of beta testing, gives other departments an opportunity to elicit

very shortly after the program ends. If you feel you are weak in any

valuable feedback, and may even help your career trajectory.

of these areas, compensate in your beta plan by requesting assistance
from the necessary group (marketing, sales, etc.) to fill in the gaps.”

87. ENGAGE WITH YOUR QUALITY TEAMS
“Make friends with your internal quality department. They will know the
product very well and will be able to tell you issues that they frequently


89. DON’T BE A FIREHOSE

encounter, which you can in turn tell the testers. They can also tell you

Beta tests generate a wide variety of feedback, covering every aspect of

what’s been fixed in a release.” -Amanda Dawson, TiVo

your product. While product managers are often interested in the big
picture, most of your colleagues only have the bandwidth for what’s
relevant to them. Providing too much information will limit their
engagement and perception of the value of your beta. Do your best to
ensure that the feedback you’re distributing is valuable to those receiving
it. This often means bugs for QA, survey results for marketing, feature
requests for product planning, etc.


90. PROMOTE THYSELF
“Market yourself internally. The job of a beta program manager is often
misunderstood and underestimated. Document what you had to do
along the way and consider presenting the entire process for recording
as a video presentation. At the very least this will give confidence to
your team that someone else could jump in and can carry on with a
rough guideline, aware of the potential pitfalls, if you were to suddenly
be abducted by aliens and not returned for at least six months.”




Closing the Beta
91. GIVE NOTICE THAT IT’S OVER

94. CREATE A SHUTDOWN PROCEDURE

Send a clear indication to testers that the project is complete. An email,

“Establish a standard process for shutting down betas — and stick to

official letter, or some other form of announcement helps bring closure

it. The details are not too important, but having a process improves

to the test and lets people understand their role is complete. It’s also

communication. We like to post a heads up that the beta will go

helpful to use this time to remind testers of the terms of the NDA should

read-only in 7 days and then be shut down completely after another

they extend past the end of the beta.

7 days. This gives customers time to file any unfinished feedback etc.”
-Simon Bosley, Autodesk

92. WAIT UNTIL YOU RECEIVE THE PRODUCT
On the other hand, if participants are supposed to ship product back to
you after testing, do not announce the project has closed until everything
is back. If they think you have officially ended up the test, they are far
less likely to return the test items. They’ll just think you’ve moved on
to other, more important matters.

93. KNOW WHO’S MAKING THE DIFFERENCE
“Cultivate good testers. Good testers are hard to find and are worth
numerous average testers. So be sure to keep track of the good testers
and make them feel appreciated.” -Amanda Dawson, TiVo



Rewarding Your Testers
95. MAKE IT WORTH THEIR WHILE

98. KEEP THE INCENTIVE SECRET

“Provide an incentive. A gift at the end of your trial, creating a better

Giving out information about incentives before or during the test is risky.

product that testers care about, or maybe inclusion in future beta

If your testers like the incentive too much, they might make up data or

programs. Testers need a good reason for spending their time testing.”

participate in other unhelpful ways to make sure they look deserving.

-Geoff Griffin, TiVo

On the other hand, if testers don’t like the incentive, they may stop
participating entirely.

96. WAIT UNTIL IT’S REALLY OVER
When testers receive their incentive for participating, that signals to

99. REMEMBER INCENTIVES SET PRECEDENT

them that the test is over. So, naturally, if you still need data or feedback

What you give as an incentive for one test affects how people think

from testers, you shouldn’t distribute incentives. Otherwise you might

about your future tests. This means you have to be careful about making

find the participation levels on those last remaining tasks a little lacking.

incentives too cheap or too expensive. If they are too cheap, testers will
feel like you do not value their input. And if you make them too expensive,

97. OFFER WHAT THEY WANTED ALL ALONG
The best incentive for a beta test is usually the product being tested.
Users signed up to test it for a reason. That’s not to say other incentives

testers will expect that from every project they test.

100. TRY SOMETHING INTERESTING

aren’t appreciated — gift cards from retailers like Amazon are usually

“Don’t be afraid to try enticing participants through techniques that

popular as well — but giving participants a non-product incentive is a

others have failed with before you or discouraged as too risky. Most great

little like depriving them of the fruits of their labor.

ideas that have received positive buzz (unique contests or methods of
feedback) were always invariably laughed at and seen as likely to fail
before they succeeded stupendously. Such ideas are rarely easy and/or
safe in planning or implementation.” -James McKey, Symantec



Final Thoughts
Hopefully, with these tips in hand, you’ll find yourself feeling more confident and energized about your next beta test. The important thing to remember,
though, is that there is always room to improve and experiment with your beta management.

We regularly publish free resources that tackle some of the topics discussed
here in more depth. And if you don’t visit our blog regularly, we hope you’ll
consider doing so. We’re always discussing a wide variety of topics aimed at
helping you tackle beta-related challenges. If you have any questions about
beta testing, you can always post your question to Quora (a social Q&A web
site that we regularly participate in), or contact us directly.

Finally, if you’d to take the next step in improving your beta testing, we’d
be happy to set up a demo or free trial of Centercode Connect, or give you
an estimate for our Managed Beta Services. You can contact us at info@
centercode.com or 800-705-6540, or sign up for a free trial of our software
at www.centercode.com/trial.



Special Thanks
At Centercode, we’re extremely appreciative of all our customers. We wouldn’t be here without them, and their feedback over the last 10 years has
been essential to shaping Connect into the industry-leading beta management platform. When we set out to create this eBook, then later decided to seek
out tips from our customers, we weren’t sure what kind of response we would get. But in typical fashion, our customers delivered for us in a big way,
providing far more than we could include in a single guide. So, we’d like to take this opportunity to directly thank all of our contributors. This eBook
wouldn’t have been the same without your support.

As stated above, not everything that was contributed made it into this edition of the eBook, but each was helpful in producing it. We plan on releasing
a second edition that will include more tips, and other contributions are welcome. If you want to be included, please comment on our blog post or send
your tips to tips@centercode.com.

100 Tips for Better Beta Tests

More Related Content

What's hot (19)

Similar to 100 Tips for Better Beta Tests (20)

More from Centercode (6)

Recently uploaded (20)

100 Tips for Better Beta Tests