Methodology
How Markish research is produced.
Research is useful only when readers can understand what was studied, how it was analyzed, and where its limits are. Markish publishes original findings only when the underlying data is legitimate to use, the method can be described clearly, and the conclusion is proportional to the evidence.
Research principles
We prioritize narrow, answerable questions over broad claims. We distinguish observation from interpretation, association from causation, and Markish analysis from third-party evidence. A report does not become stronger by omitting uncertainty.
Data sources
Depending on the study, eligible sources may include aggregated product data, anonymized workflow or content-performance data, publishing activity, legitimate survey responses, manually assembled research datasets, and public market data. Each report identifies the source actually used; this page does not imply that every source is available for every study.
Data preparation
Before analysis, Markish documents the unit of analysis, time period, inclusion and exclusion criteria, and aggregation approach. Relevant checks can include duplicate records, missing values, invalid timestamps, test or demo accounts, incomplete workflows, outliers, and system activity. A report states material exclusions instead of hiding them.
Metric definitions
Every report defines the metrics it uses and, where useful, gives the calculation. For example, an approval-duration measure must specify the exact start and end timestamps used. Definitions are report-specific: Markish does not assume that terms such as engagement rate or publishing consistency have one universal formula.
Analysis and benchmarking
We compare like with like where possible and preserve context such as brand objective, platform, audience, and time period. Benchmarks are published only when the dataset and comparison logic support them. Observational analysis may identify patterns or associations; it does not establish causation without a method capable of doing so.
Quality checks and review
Before publication, at least one reviewer checks the dataset, calculations, interpretation, wording, charts, limitations, and product references. The team should be able to trace a major finding to its source data, cleaning decisions, metric definition, exclusions, and analysis method.
Privacy and aggregation
Markish research must not expose client identities, private account information, personal information, private messages, confidential brand data, or individual user behavior without appropriate authorization. We prefer aggregated or anonymized reporting and exclude information that creates a meaningful re-identification risk.
Limitations
Every report includes its meaningful limitations. These may include sample composition, platform-data differences, short analysis periods, seasonality, missing context, or the fact that correlation does not demonstrate cause. Readers should be able to see what a result does and does not support.
Updates and corrections
If a material error is identified, Markish verifies it, corrects the publication, updates the date, and documents the material change. When a substantially expanded dataset changes the meaning of prior findings, Markish should consider a new version or report rather than silently rewriting history.