Content Strategy & Research

Original Research Planning

Design the question, sample, analysis, and safeguards before collecting any data.

Original Research Planning defines the question, method, sample, analysis, and ethics protocol before data collection begins. The plan gives every later claim a study design that can withstand scrutiny.

You walk away with a signed-off research plan, a pre-specified analysis, and a publication brief that is honest about what the study can and cannot claim.

Design a defensible study
Researcher sketching a study design of question, sample, method, and analysis plan on a whiteboard

Some of the 500+ brands we've worked with

See all references
  • EY
  • D&R
  • Abdi İbrahim
  • Mynet
  • Silverline
  • Sporjinal
  • Axa Hayat Emeklilik

How we run it

How the study gets designed

The research question sets the method, sample, safeguards, and analysis plan. We settle each of those decisions before fieldwork can influence them.

  1. Pin down the actual research question

    We separate the real question from the conclusion someone's hoping for, and check it against existing literature so it isn't answering something already settled.

    A defined, defensible research question.

    AI assist
    Agents can review the existing literature and show which questions have already been answered and how. Narrowing to a question your data can actually answer is a researcher's decision.
    Human gate
    The requesting stakeholder confirms this is the real question, and that the answer isn't already decided.

  2. Choose a method the data can support

    Sample size, source, and analysis method get matched to what the question actually requires. Cheapest and fastest are separate considerations.

    A method and sampling plan sized to the question.

    AI assist
    Agents can compare candidate methods against the sample you can realistically reach and show the tradeoffs. Choosing a method the data can support, and accepting its limits, is a human call.
    Human gate
    A researcher signs off that the sample size and method actually fit the question.

  3. Build the ethics and consent protocol

    We settle the privacy requirements, final consent language, and disclosure obligations before participant outreach starts.

    An ethics and consent protocol ready for legal or ethics review.

    AI assist
    Agents can draft the consent language and the data-handling steps from your existing policy. Your legal or ethics reviewer approves consent and privacy before any participant is contacted.
    Human gate
    The first participant cannot be contacted until legal or ethics has approved the final consent language.

  4. Specify the analysis in advance

    How the data will be analyzed gets written down before collection starts, so results can't quietly get reshaped to fit a preferred story.

    A locked analysis specification.

    AI assist
    Agents can write the analysis specification into a pre-registered form and flag comparisons that would be exploratory. Fixing the analysis in advance is what stops the result from being chosen afterwards.
    Human gate
    The researcher locks the analysis plan in writing before any data comes in, so it can't shift later.

  5. Assign roles and timeline

    Research, legal, editorial, and communications owners agree on who does what and by when, so the study doesn't stall on an unclear handoff.

    A staffed timeline with named owners.

    AI assist
    Agents can build the timeline from the agreed steps and show where ethics review sits on the critical path. Assigning the roles is done with the people who will hold them.
    Human gate
    Research, legal, editorial, and communications owners each confirm their own slot in the timeline.

  6. Write the publication brief

    We define how the findings will eventually be written up and released, including what limitations must be disclosed alongside the results.

    A publication brief ready for the writing and PR teams.

    AI assist
    Agents can draft the publication brief from the design, including the claims the study will not support. Recording that honest findings may contradict the preferred narrative is a decision we make out loud.
    Human gate
    The researcher confirms every limitation that must travel with the findings is named in the brief.

The study design needs a person to stand behind it

Literature searches and competing claims are good work for agents, and an early design draft can help the team see what still needs deciding. The final calls are human ones. A Zeo researcher is accountable for the sample and method, while your legal or ethics reviewer must approve consent and privacy before participant outreach starts. The plan also records the possibility that honest findings may contradict the preferred narrative. We will not design around a predetermined answer.

What you get

The pre-fieldwork research plan

You receive an approved study plan that prevents design decisions from being improvised after fieldwork starts.

  • a content brief with checklist marks

    Research plan

    Question, method, sample and source plan, and ethics protocol, ready for legal or ethics sign-off.

  • a measurement note with a small trend line

    Analysis specification

    Exactly how the data will be analyzed, locked before collection starts so it can't shift to fit a convenient story.

  • a style guide with marked rules

    Publication brief

    How the eventual findings should be written up, including the limitations that need to travel with the results.

We call it done when: research, legal or ethics, and communications owners have all signed off on the plan, and the publication brief is honest about what the study can and cannot claim.

Who it fits

The question deserves a real study

This method designs the study. Fieldwork, analysis, and report writing happen after the plan is approved.

A good fit when

  • Your thought-leadership idea needs original data, but every number must survive a skeptical reader checking the sample and method.
  • The team has a real question whose answer is still open. A feasible method can collect evidence that speaks to it.
  • Participant outreach touches consent, privacy, or sensitive data, so legal or ethics approval must be settled before collection begins.

Better handled as other work when

  • Leadership has already chosen the finding, while the study is expected to produce numbers that support it regardless of what the data shows.
  • The only reachable sample is small or biased, but the publication still needs to present it as representative of people the study never reached.
  • The results need to appear without the method or limitations, so a reader would have no way to judge how far the findings can travel.

People who write and ship this content

Some of these consultants write the pages themselves. Others plan distribution or check how a piece performs once it's live, so the team below covers more than one stage of the work.

Content we've produced on this topic

Articleby Didem HimmetliMaximizing the Power of AI Assistants in Data AnalysisArticleby Samet ÖzsüleymanMaking Efficient Decisions with Correlation Analysis in SEOArticleby Samet ÖzsüleymanPredictions and Survey Results on the Future of Generative AI

Tools we use

Tools behind this work

  • SurveyMonkey

    runs the population, confidence, and margin-of-error math behind the required sample size

  • Jupyter

    the pre-registered analysis plan, run against real data in the same notebook

  • Airtable

    roles, consent status, and the timeline, tracked as one linked base

  • Notion

    the consent and ethics protocol, open for legal's direct comments

Start the work

Have a research question that needs a defensible study?

Bring us the question, even if the answer may be inconvenient. We'll work out the study design and the claims it can responsibly support.
Talk to a Content Marketing specialist
Two people shaking hands on the start of the work

Before we start

Frequently Asked Questions

A result guaranteed to support the narrative would make the study something other than research. We design for a defensible answer. That includes the possibility that it contradicts the hoped-for story. Hiding that possibility would put the brand's credibility at risk as soon as someone checked the methodology.