What the exam actually feels like

This exam is much more about authoring decisions inside Cognos Analytics V11 than about administration, architecture, or random product trivia. Expect standard multiple-choice and multiple-response questions rather than hands-on labs. The pressure comes from scenario wording: IBM likes to describe a reporting requirement in business language and then ask which authoring feature satisfies it with the fewest steps or the most appropriate object.

A practical way to pace it is to assume roughly a little over a minute per question if you get an exam form around the usual IBM certification length. That matters because some items are fast definition checks, while others are mini-scenarios comparing list, crosstab, chart, dashboard behavior, drill-through, bursting, prompt pages, and active report interactions. Don’t spend equal time on every item. The time sink questions are usually the ones that give you a report requirement and then offer four options that all sound technically possible in Cognos.

What stood out to me:

IBM often tests the best authoring method, not merely a method that could work.
The exam assumes you know the difference between Report, Dashboard, Story, Exploration, and Data Module at a practical level, but this exam leans heavily toward report authoring behavior.
Questions often hinge on whether something should be done with a query calculation, layout object, conditional style, prompt, drill-through definition, bursting definition, or report property.
You can usually go back and review in IBM exam delivery, so mark the long scenario items and keep moving if you’re stuck between two plausible authoring choices.

If you’re preparing from practice questions, don’t just memorize right answers. Rebuild the scenario mentally in Cognos Authoring: “Would I solve this in the query? in the layout? as a prompt? with a master-detail relationship? or with a drill-through target?” That is exactly how the real questions are framed.

Where the points really are: authoring core, prompts, and output behavior

The exam is not evenly spread across every Cognos feature. The highest-value study area is the cluster of topics around building reports, choosing report objects, filtering/grouping/sorting/aggregation, and controlling output behavior. If your revision time is limited, spend most of it there.

The concentration areas I saw repeatedly were:

Report object selection
- When to use a list versus a crosstab versus a chart.
- When a requirement needs a repeater, table, or singleton rather than another list.
- How sections change output and readability.
Queries and calculations
- Detail filters vs summary filters.
- Query calculations vs layout-level expressions.
- Aggregates that look simple but behave differently in grouped reports.
Prompts and parameter use
- Prompt pages, prompt controls, default values, cascading prompts, and parameter mapping.
- The difference between filtering data in the query and just collecting user input for later use.
Conditional formatting and conditional blocks
- IBM likes to test the distinction between changing style and changing visibility.
Drill-through, drill-up/down, and master-detail relationships
- These are easy to confuse if you only know the feature names.
Bursting, scheduling, and output formats
- Especially when the question mentions distributing personalized output to recipients.
Active reports / interactive behavior
- Not every exam form hits this hard, but enough do that you should know what active report controls are for and when they work offline.

A trap here: some candidates over-study framework modeling and administration. You do need enough understanding to know where an author stops and a modeler/admin begins, but this exam is about the author’s choices in Report Authoring. If a question asks how to make a report show different data to different recipients automatically, think bursting, not security classes or package redesign.

The list-crosstab-chart questions are more subtle than they look

IBM loves asking for the “most appropriate” report object, and this is where people lose easy points because they think in terms of what can be forced to work. On this exam, answer based on what an experienced Cognos author would choose first.

Use this mental shortcut:

Choose a list when the requirement is row-oriented detail, grouped detail, or a simple summary with repeatable records.
Choose a crosstab when the requirement is dimensional comparison across intersections, especially categories on one axis and measures across another.
Choose a chart when the requirement emphasizes trend, comparison, composition, or visual analysis rather than row-level reading.
Choose a repeater/repeater table when you need the same object repeated for each value, often for invoice-style or card-style output.
Choose a singleton when the requirement is a single value such as a title metric, date, selected prompt value, or KPI-like display.

The gotcha is that IBM often wraps this in business wording. Example patterns:

“Users need to compare revenue by product line across quarter and region.” That is usually crosstab or chart depending on whether the emphasis is exact intersection values or visual comparison.
“Users need invoice details for each customer, with a page break per customer.” That points toward list/report grouped by customer, possibly with sections or page set behavior, not a dashboard.
“Management wants one page showing current year sales total and prior year sales total.” That smells like singletons or a small summary report, not a full list.

Another common trap is aggregation in lists. If the report is grouped and the question asks why totals look duplicated or wrong, think about:

whether the data item is aggregated at the correct query level,
whether the grouping creates repeated summary values,
whether a summary should be placed in a group footer rather than the detail row,
whether the expression should use a proper aggregate like total() at the intended scope.

Don’t answer these from SQL instinct alone. IBM wants the Cognos authoring fix, which is often about report structure and object placement as much as expression logic.

Prompts, parameters, and filters: this is where IBM hides the traps

A lot of the exam’s trickiest questions live here because the wrong options are all half-right. You need to separate four things clearly: prompt controls, parameter values, query filters, and layout expressions.

What to remember:

A prompt collects input.
A parameter stores or passes that input.
A filter limits returned data.
A conditional layout/style expression changes what the user sees, but may not reduce the dataset.

IBM likes scenario questions such as: a user wants to choose a region, then only products in that region should appear in the next prompt. That is a cascading prompt pattern, not just two unrelated prompts on the same page. If the second prompt must depend on the first, look for the option involving parameterized filtering of the second prompt query.

Specific gotchas worth drilling:

Detail filter vs summary filter: if the requirement says “return only product lines whose total revenue exceeds X,” that is not a simple detail filter on individual rows. The filter must align with the aggregation level.
Optional prompts: IBM may ask how to let users run a report without selecting a value. Know the difference between making a prompt optional and writing filter logic that handles null/no selection cases.
Prompt page vs generated prompt page: if the question is about control and customization, manually authored prompt pages are usually the better answer.
Prompt values in titles or headers: this often uses parameter display/value handling, not a query calculation.
Use of static choice prompts vs value prompts: if the values come from fixed business options, static can be valid; if they must reflect package data, value prompts are more appropriate.

One pattern I saw more than once in practice sets: IBM describes a requirement that sounds like formatting but is actually filtering. Example: “Show only top-performing sales reps in red.” If the requirement is just to visually identify them, that’s conditional style. If the report should exclude all others, that’s a filter. Read the requirement as if you’re implementing it in the properties pane.

If you’re shaky on prompt macros or advanced parameter behavior, at least get solid on the common author-level use cases: default selections, prompt relationships, passing prompt values into drill-through, and using prompt values in report expressions.

Drill-through, master-detail, and active report behavior

These features get mixed up constantly, and IBM knows it. The exam will often present a requirement that could be solved by more than one interactive technique, but only one is the clean Cognos answer.

Here’s the distinction that helped me:

Drill-through: navigate from one report to another target report, often passing parameter values.
Drill up/down: navigate within dimensional hierarchies when the package supports it.
Master-detail relationship: coordinate objects in the same report so one object controls the data shown in another.
Active report controls: provide interactive behavior, often intended for disconnected/offline consumption scenarios.

If the question says users click a summary value and open a separate detailed report filtered to that value, think drill-through definition. If the question says selecting a chart bar should update a list on the same page, that’s much closer to master-detail or an active report interaction depending on the exact wording.

For drill-through questions, know these pain points:

Passing the correct parameter from source to target.
Matching data types and parameter names logically.
Deciding whether to open in the same window, new window, or another format when the requirement mentions user experience.

For active reports, don’t overcomplicate it. IBM usually tests the core idea: interactive report content packaged for users, often with controls like buttons or selections that work without constant server interaction in the same way as standard prompts. If the wording mentions mobile/offline-style interactivity, active report should be on your radar.

A subtle trap: if the requirement is really just “click to a more detailed view,” don’t choose active report because it sounds more advanced. IBM usually rewards the simpler native feature that directly matches the requirement.

Bursting, scheduling, and distribution questions are easy points if you know the vocabulary

This exam likes operational reporting scenarios: monthly statements, manager-specific reports, region heads receiving only their own data, or a single report definition sent to many recipients. That is classic bursting territory.

What to lock in:

Bursting splits report output based on a burst key and distributes it to recipients.
It is the right answer when one report must generate personalized slices for different consumers automatically.
If the question mentions email delivery, recipient lists, or scheduled distribution of individualized output, bursting should be your first thought.

The exam may contrast bursting with:

running separate reports manually,
using security alone,
adding prompts for each recipient,
creating multiple report copies.

Those are usually wrong when the requirement is automated personalized delivery.

Also be clear on output format implications. IBM may ask which format preserves a certain type of interactivity or which is best for downstream distribution. You don’t need to memorize every format edge case, but know the broad differences between HTML, PDF, Excel/CSV-style outputs, and active report outputs. If the requirement is static, printable, and consistent layout, PDF is often the practical answer. If it emphasizes preserving rows for analysis/export, a spreadsheet-oriented output may make more sense.

One more trap: scheduling and bursting are related but not identical. Scheduling defines when it runs; bursting defines how output is partitioned and distributed. IBM sometimes writes answer options that blur those on purpose.

What looks easy but burns people: conditional formatting, calculations, and query reuse

The questions that look beginner-level are often where IBM slips in precision traps. Three areas deserve more respect than people give them.

First, conditional formatting. You need to know whether the requirement is to:

change text color/background/style,
hide/show an object,
switch entire sections of layout,
or return different values in an expression.

These are not interchangeable. If the requirement says “display one chart for executives and another for sales managers,” that sounds more like conditional blocks/layout control than simple conditional style.

Second, calculations. Be careful with percentages, running totals, and aggregate calculations in grouped reports. IBM can ask why a calculation is wrong, and the answer is often one of these:

the expression is calculated at the wrong query level,
the aggregate scope is wrong,
the report object placement causes repeated values,
or a summary function is applied to already summarized data.

Third, query reuse and multiple queries in one report. If a report uses a chart and a list or combines unrelated objects, be alert to whether they should share a query or use separate ones. If the requirement implies different granularity or different filtering logic, separate queries may be necessary. IBM doesn’t usually make this ultra-technical, but it does expect you to recognize when one query design causes incorrect output across different objects.

A useful prep exercise is to take one sample report requirement and ask five implementation questions:

What report object is best?
What query or queries are needed?
Where should the filter be applied?
Is the requirement about data restriction or display behavior?
Does navigation happen within the report or to another report?

That mirrors the exam better than passive reading.

My final-week cram list for C2090-621

If I had to compress the last stretch of prep into a blunt checklist, it would be this.

Rebuild these topics until you can explain them without notes:

Difference between list, crosstab, chart, repeater, singleton, and table.
How to implement grouping, sorting, summaries, and sectioning in a report.
Detail vs summary filters with examples.
Prompt types, cascading prompts, optional prompts, and parameter passing.
Conditional styles versus conditional blocks/visibility.
Drill-through setup and when it beats master-detail.
Master-detail relationships inside a report.
Bursting mechanics and when a scenario clearly calls for it.
Output format choices for print, export, and interactive use.
Basic active report purpose and controls.

Use the 77 practice questions smartly, not just repeatedly:

Tag every miss as one of these buckets: object choice, query logic, prompt/filter confusion, navigation/distribution, or formatting/layout.
If you miss a question because two answers both seem possible, write down why IBM prefers one. That’s the real learning.
Re-answer the scenario questions from memory in implementation language: “I would add a value prompt, map it to a parameter, apply a filter to the query, then display the selection in a title.”

The night-before topics worth a quick refresh:

authoring interface terms and object names,
common report properties,
drill-through parameter behavior,
bursting terminology,
active report basics,
and examples of when a requirement is solved in the query versus in the layout.

The biggest mistake I see with this exam is studying Cognos too broadly. For C2090-621, keep asking: What would an IBM Cognos Analytics Author actually do in the report authoring tool to satisfy this requirement cleanly? If you answer from that perspective, a lot of the trick questions stop being tricky.