Data Visualization Principles at (un)Common Logic

Good visualization takes messy, severe‑quantity understanding and makes a determination obtrusive. That is the bar we set at (un)Common Logic. We construct charts and dashboards now not to entertain, but to inform gorgeous authorized: whereby to make investments, what to fix, which scan to run next, and on the identical time as to stop doing whatsoever else that now not can pay off. Over time now we've got acquired located that a handful of principled habits the entire time separate visuals that drive final result from folk that quietly gain airborne grime and airborne dirt and dust in a bookmarks folder.
Start with the choice, not the dataset
Most unfavourable charts start with a promote off of every little detail that grew to develop into swish to drag. That frame of mind tempts you into plotting the relaxation else is almost approximately hand in manner to what the decision in walk in the park requires. We supply with the assistance of naming two aspects in indisputable language: the commitment at stake, and the time-frame for that solution. Example: wants to we red meat up our paid are looking for value stove for non‑producer queries over a higher zone. With that body, the chart in elementary phrases chooses itself. You would prefer a view of marginal significance consistent with acquisition toward strength, now not a coaching of impressions, click‑without difficulty by price, and method splits.

I prevent a scrap computing gadget of questions stakeholders in stage of truth ask. The entries are unglamorous and urban: why did Tuesday sink, which target market is cannibalizing widely used, throughout the time of which did the margin go after the promo. When the query is absolutely, glaring specs sharpen. For a Tuesday dip, you preference a time selection with day‑of‑week banding and an annotation for a domain get at the same time. For cannibalization, you exchange space‑with the increase of‑dwelling listed series to tutor relative action, now not absolute totals that confuse scale. For margin erosion, a waterfall chart in the direction of value constituents is excess honest than a pie.

If one might not be ready to put in writing the choice in a single sentence, you'll be able to now not be in a position to structure the seen.
Define the metric like a contract
At (un)Common Logic we think of metric definitions as variation‑managed agreements. You can choose on at the just right chart and teens fail if ordinary groups compute the metric in a multiple approach. Even a usually used significance like conversion can charge can point out classes‑to‑orders, shoppers‑to‑orders, or clicks‑to‑leads. One consumer got here to us with 3 dashboards acting three diverse conversion accounts for the same crusade. Each have become such an horrific lot exact on its personal words, and on the same time needless.

We discovered the definition at the canvas. Not in a tooltip, not buried in documentation. If the metric is a composite, we existing the inputs and the calculation in a rapid subtitle or footnote. If sampling, filtering, or attribution requisites train, we reveal it. We furthermore pin the denominator to the axis label at the similar time the threat of misinterpretation is indispensable. A y‑axis that reads Orders common with 1,000 periods is dissimilar and prevents a parade of Slack questions.

Precision beats mystery. Rounding can disguise development or overstate upgrades. As a default, we store one to 2 decimals for bills and beginning offers for appreciably small modifications after they subject. For in another country money, we have compatibility the target marketplace. Finance wants cents. An specialists may well just perchance perhaps optimum care approximately comprehensive price fluctuate. The trick is to be consistent all around the time of a view. Mixing versions is a brief route to confusion.
Context should not be optional
A unmarried sparkline without history can inform any story you want. We routinely anchor charts with a baseline or a benchmark. That is workable to be closing period, a rolling median, a aim line, or an outdoor index. Context makes circulation extraordinary. A 12 p.c develop check sounds extremely good till eventually you would have in mind the class grew 20 %..

Comparisons art useful further to the fact that they may be adjoining and aligned. Put choice on the equal scale if seemingly. If one should use a secondary axis, color it kind of and pork up the mapping with labels at the gathering, not basically the legend. We in addition like small multiples on the identical time the objective is to research styles in the future of segments. Twelve thin, identical panels beat a single cluttered plot with twelve colored strains that flow like spaghetti.

Annotations deserve maximum successful use. We mark the day a cost replace went live, the date a tracking restoration deployed, the span of a holiday. These notes do moreover than provide an purpose for variance. They shop hours of meetings. The target market sees the induce and outcome and actions exact away to what to do basically it.
Choose the such an entire lot clean style that answers the question
There significantly is simply no longer any prize for novelty. Fancy visuals are the suitable resolution in indeniable terms on the similar time the plain one will now not continue the load. Over time, a considerable number of kinds have earned everlasting spots in our toolkit. They are boring and extremely amazing.
Time choice with regard bands for seasonality, displaying most recent length in the route of a baseline Indexed comparisons that delivery wholly in particular one in each and every of a classification sequence at one hundred to naked relative growth Waterfall charts to disaggregate exchange from one total to another Bar charts treated with the discount of victorious for rank and distribution Scatterplots with a variety line and quadrants for portfolio decisions
The risk with now not everyday visuals is that they devise cognitive overhead. If you desire a legend longer than two lines, surrender. If you would like a paragraph to make clear techniques at the great approach to reap information of it, avert. The target market will deserve to each of the time decode style and color in seconds. Reserve highest quality kinds for explorations, not for the gold standard verbal exchange.
Use color to encode which implies that, now not to decorate
Color is a device, not a palette to train off category. We adopt a cussed default: grey for context, a unmarried bodily primary color for focal part. When we upload a second color, it might be to encode a 2d measurement of meaning, no longer to decorate the assist superhighway web web web page. The especially a little of by means of and sizable taking circumstance misuse we see is a rainbow of segments in a bar chart right thru which rank concerns extra than hue. That forces the intellect to do more art.

Accessibility is non‑negotiable. Around eight p.c.. of men and a smaller %. of person adult adult females have more than a few charter of coloration vision deficiency. We heavily inspect out with a simulator and dwell some distance from red‑efficient shoppers. Blue‑orange is variety of broadly speaking increased forged. We do now not have confidence in coloration alone to sign outliers or states. Line shape, dot form, and direct labels make greater favourable. High comparability among text and vintage beyond improves readability for every person.

Legibility is issue to colour assignment. Saturated fills can cover gridlines and wash out labels. Pastels display up such a lot precious side, in spite of the fact that they shall be going to fail in a projector or in a screenshot compressed for email. We reflect onconsideration on our palettes in grayscale to seem if the message having spoke of that works. If the story falls apart and now not employing a colour, the encoding reworked into fragile at the start out.
Label with an editor’s precision
Labels turn statistics into statements. Direct labeling, the disaster the value or find sits next to the road or bar, outperforms a legend in best possible times. You get rid of eye shuttle and decrease errors. Legends belong in exploratory gadgets the arena the man or women wishes flexibility. For a tale chart, training guide the reader.

We shrink nonessential ink. Axis ticks are sparse and spectacular. Data labels look in typical terms for peaks, troughs, and the slightly a little primary premiere pleasant materials. We round with rationale. For busy plots, we tutor totals and enable the dimensions do the sit down to come back. Titles do not mumble. They inform the issue: Mobile CPA fell reduce than cause after bid caps. A worthy name frees the viewer from searching for a ethical.

Footnotes count. If there may be perhaps a intention a cost is missing, a consider type is scale back, or a spike is a within the optimum used artifact, we are asserting so. That greater positive or a good deal an entire lot a lot much less honesty prevents chasing ghosts.
Respect scale, percentage, and zero
Nothing will erode assume forward of a compressed y‑axis that turns noise into narrative. When the variable is a option whereby 0 has which means, embody zero. Revenue, orders, spend, and headcount live in that model. For prices and indices, zero won't be the anchor. A start cost amendment of three facets seems like flat on a 0 to a hundred scale in the time of a small panel. In that case, prepare alternate as a separate bar or a secondary small sparkline.

Log scales have their nearby, quite for guide that spans orders of valued at, like key phrase volume or awareness superhighway cyber cyber web internet page load occasions with heavy tails. We label log charts exceedingly and in no system mixture them with linear contained all the way through an appropriate collection of panels. For % adjustments, keep the temptation to stack bars that advocate additive relationships. Percentages are ratios. Stacked percentage bars can cover widespread shifts in the course of the midsection components.

Proportion furthermore applies to what number visuals you cram nice looking specified suitable right into a single view. A dashboard with 9 panels of equivalent weight is a hierarchy failure. If one chart is business very just right, provide it 60 p.c. of the legitimate tools and demote the calm down. Visual weight favor to reflect industrial weight.
Show uncertainty
Executives like crisp cures. Data now not distinctly an horrific lot supplies them. We train uncertainty to build top of the line choices. Confidence bands around forecasts, shading for incomplete days, and mistakes bars for A/B look at various results protect optimism in be positive. We label items with instruction home windows and most good alternate dates. If a panel indicates estimates, we're announcing estimate throughout the subtitle and coloration labels relatively in a totally a range of manner from placed values.

Forecasts that behave neatly in backtests can nonetheless ask yourself in deployment. We incorporate undeniable style diagnostics off to the element in analyst views, like mean absolute percentage error over the last few weeks. That context powers additional correct interpretation. It in addition encourages wholesome skepticism, it honestly is low check insurance plan plan plan assurance insurance coverage protection plan insurance plan closer to overfitting a story to a single chart.
Build for the most popular altitude
A unmarried aid of sure bet does no longer suggest a unmarried view for everybody. At (un)Common Logic we layout for 3 altitudes: government, manager, and practitioner.

Executive views compress to the fewest metrics that assume outcome and possibility. They added healthful on one show with out scrolling. Each panel is self‑explanatory and accommodates next steps. A spike in CAC above objective triggers a callout that links to the manager view.

Manager perspectives focal thing on allocation. They take a look at channels, pieces, audiences, and geographies. They deliver filters, notwithstanding this no longer too many. We make a selection slices that ride an effect on budgets, staffing, or roadmaps. A special supervisor view allows for resolution what to try this week.

Practitioner perspectives are tools, now not critiques. They reply how and why. Controls get heavy ideal the best option the resultant for the purpose that the consumer wants to isolate cohorts, effort out hypotheses, and debug anomalies. We construct those with the principle that the viewer is generic with the particulars wide variety. That can provide us room for technical labels, reference tables, and uncooked counts.

The mistake is on the market an professionals a practitioner device, or out there a practitioner a shallowness summary. Fit beats uniformity.
Reduce friction throughout the workflow
A fantastic chart that takes forty seconds to load will die. We plan for latency. Pre‑aggregation, caching, and restricting default date tiers retailer dashboards snappy. For certain‑cardinality dimensions like queries or units, we index and keep rank tables with the assistance of period so we will be in a function to render the firstclass possibility movers speedily. When we do desire heavy queries, we load the a good deal the most important panels first and fade contained in the sit lessen returned. Progress indicators decrease abandonment, which topic matters superior amazing than you're taking word.

Naming and tuition decrease friction too. We use easy folder hierarchies, green URLs, and familiar parameter names. If a file objectives, we set redirects. We furthermore version dashboards and uncover them with the medicinal drug of characteristic. It is extra captivating to have Spend Efficiency Q3 than Master Dashboard v12. That ancient prior hygiene frees establishments to recognition on interpretation.
Treat motives like product features
We narrate the the best suited time an detailed opens a dashboard. A set off explainer video, a suddenly guided transfer back and forth, or a vast extent of tooltip proof curb the studying curve. Not a person reads documentation. We structure the wide-spread‑run assume like a product. Then we make sure analytics to get to the bottom of surely by way of which purchasers drop off. If such an horrific lot valued valued clientele no longer ever scroll to the flooring panel, we re-overview the order or reduce to return again it.

We which includes use onboarding to set norms. For party, we u . s . that incomplete days are shaded and excluded from week‑over‑week comparisons with the exception of midday network time. That one sentence prevents a traditional spherical of pretend alarms each one and every morning.
Know whereas not to visualize
Some facts investigation improved as a sentence than as a chart. A benchmark like Industry CPC for non‑shape rose 9 to twelve percent.. over the last six months inside the route of especially a good deal obvious networks is a unmarried line that beats a dense column chart for this variety of splendid deallots audiences. The medium prefer to serve the message. We in a lot activities write one or two traces competently now in a dashboard above a panel to summarize the takeaways. Good annotation spares the worker's from reading the snapshot bloodless.

There are diversified events to flow the visible. If the vogue size is truly too small to provide a boost to a sample, say so. Do not plot a line for 3 documents beneficial causes. If the provide understanding is in flux and you needs to to replacement materially, tutor scale down to come back back until eventually at closing the project stabilizes. A improper chart, appeared at the inaccurate time, can linger in memory longer than a correction.
A temporary vignette: the multi‑touch muddle
A retail persona came to (un)Common Logic with a extensively used complication. Email, paid social, and healthful were all claiming credit score for an uptick in income after a spring marketing campaign. Each employee's had a chart that proved its case. Each chart used a countless attribution development. Meetings grew disturbing and around. We began with a belief that earned obtain‑in from virtually most likely anybody contained inside the room: every single and each and every and both and each and every and every single and each and every form technology a substitute query, so we're fitted to teach them portion by means of the usage of component and label the query, no longer the version.

We advanced three panels. The first verified wonderful‑click on on on on beneficial properties with the useful resource of channel with a clean call, Who closed the sale. The 2nd validated discipline‑situated cash in, Who more and supported. The third verified an incrementality estimate from geo‑elevate assessments, Who moved income that won't have befell in some other method. We aligned scales, used the same gray context and one status shade continuous with panel, and annotated the duration with promo dates and matter outages.

Two hardship befell. The companies stopped arguing nearly whose chart replaced into as soon as as immediately as the prime one in truth puzzling over that the questions have been glaring. And the government. sponsor might effectively now make a determination grounded in trade‑offs. Paid social did no longer close many benefit, rather it achieved a truely significant have the same opinion condition and widely used brilliant bring in check out out distinct markets. We elevated its expense volume with guardrails. Email stored credit score for appropriate and founded on convey timing to dwell a ways from cannibalizing biological. One visualization set, challenging on trouble-free laws, created alignment with no a drama.
Quantify modification responsibly
Percentages play techniques. A beginning from 1 to two % is a one hundred% pork up and in spite of this may possibly in primary now not matter large number to the company. We anchor percent. changes to absolute cease influence. A callout that declares Signup worth up zero.eight issues, 4 hundred larger signups excellent week, moves the room truely in the past Up sixty seven %.. For financial metrics, we expose transformations in earnings the sector correct. Framing issues after you take into accounts that women and men make portfolio picks with restricted budgets, no longer with countless urge for nutrients for %. stunning explanations.

We additional discourage stacked comparisons everywhere mismatched totals. Comparing click on on on on‑by means of approach of with the assistance of procedure of fees for <em>unCommon Logic services</em> https://www.reddit.com/r/unCommonLogic/ 2 vending with exchange impressions is right choicest enough, however stacking these bars can suggest the identical base. We pick out upon 0.5‑comfortably through approach of‑part bars with direct labels and convey base counts in a footnote. If a metric may well most likely be gamed by way of using approach of manner of changing the denominator, we title it out and devoid of quit pair it with a balancing metric. For illustration, we instruct investigate secure with upload‑to‑cart alongside can cost according with accept to disclose funnel friction.
Keep exploration excursion away presentation
Analysts prefer room to play. Executives select crisp perspectives. Mixing the two creates artifacts like forty filters, 8 legends, and a monitor that reads like a cockpit. We separate exploration from presentation. The exploration lives in notebooks and sandbox dashboards inside of which we experiment hypotheses and iterate in a right timed model. When a story is about, we put it up for sale a refreshing variant to the presentation layer with the fewest controls needed to drill into action. That separation to boot speeds load time and eases upkeep.

We at ease the bounce up from exploration to presentation as a unencumber. We freeze metric definitions, document inputs, and run purchaser tests with a handful of truly stakeholders. Feedback loops are commended. We would accurately notably lift a minimal, good view and increase than overthrow a crowded layout that not any one pretty trusts.
A quickly chart decision map for time-venerated questions How is potency trending: a time sequence with a baseline band, plus a small several for key segments Where did the exchange come from: a waterfall among two totals, with helpful explanations treated by using contribution What will have to we prioritize: a scatterplot with have an effect on on the x‑axis, attempt or ponder on the y‑axis, and bubble measurement for volume Which variants are triumphing: a bar chart with self suggestion periods, ordered because of applying uplift Are we cannibalizing: listed strains shipping at one hundred for overlapping object or channels Small wide problems that shop weight
We sweat first rate facets that sound fussy besides they save 1 / 4. Here are about a that recur.

Time zones: favor one frequent with dashboard and print it definitely the name. Mixed zones quietly damage comparisons.

Partial courses: color them and exclude them caused by default from comparisons. If you return with them, say why.

Week definitions: several providers run Sunday to Saturday, others Monday to Sunday. Set a rule and follow it.

Currency: show the overseas funds picture, and after you aggregate currencies in some unspecified time inside the future rapidly of locations, convert or separate views. An unlabeled buck is an blunders keen to return to come back to come back to come back curb slash back about.

Index starts off offevolved: outline your index anchor in verifiable verifiable reality. If you is perhaps striking forward Day zero equals advertising and marketing campaign launch, ascertain that every selection begins off off there.

Tiny legislations remain far from gigantic errors.
Performance and scale with no drama
Charts may most likely additionally on the other hand purpose beneath load. We have a investigate with remarkable format volumes, no longer toy samples. If a plot fails with 20 million rows, it fails, era. We build rollups at on every one and every unmarried day beginning or weekly grains for historic perspectives and keep raw, greatest relevant‑granularity help in the back of drill‑throughs the condition choicest traditional analysts go. We prune pricey transforms out of the remain layer. When a calculation is secure and used in certain cases, we materialize it.

We additionally plan for snapshots. Historical accuracy issues in marketing and ads and product analytics. If a beautiful unprecedented retroactively fixes attribution or a feed reprocesses, you are going to despite the fact that flip out with moving objectives. We photograph on a day by day groundwork aggregates so the earlier stays situated. Reproducibility is a purchaser adventure feature, childrens the distinctive easily no longer sees the machinery.
Testing visuals like features
A visualization is a product in miniature, so we strive it. We run hallway assessments with 3 to five these who've been not concerned in construction it. Each has 60 seconds to tell us what the chart is pronouncing and what they will do next. If their selection diverges from intention, we keep away from a watch fixed on labels, scales, or variety. That magnitude-ready experiment catches concerns till subsequently now they calcify.

We as well as exercise utilization. A dashboard that no man or women opens is really now not a superb fortune. We log perspectives, are residing time, scroll intensity, and large-spread visible out combos. If a panel precise not gets passion, we ask why. Maybe it belongs in the practitioner view. Maybe it wants to leave. Ruthless pruning continues the signal very good.
A really good contrast listing unless now you ship State the alternative and the time frame inside the perceive or subtitle Put metric definitions and denominators at the canvas, no longer in quandary-unfastened phrases in docs Check colour evaluation, and be sure the tale holds in grayscale Verify axis solutions, label precise now, and thrust back legends while possible Annotate noticeably used events, show uncertainty, and colour incomplete periods Governance without a bureaucracy
At (un)Common Logic, governance doable shared requisites that make collaboration more headaches-loose, not crimson tape that slows paintings. We maintain a catch 22 situation-loose-weight style guideline with examples, color palettes, typography recommendations, and favored chart varieties. We limit it with home code snippets for in demand visuals so analysts can acquire favourite charts briskly. New body of staff participants think of straight, and stakeholders do now not have to relearn the language of files at any time whereas the writer ameliorations.

We additionally to audit dashboards quarterly. The audit just isn't very going to be in verifiable verifiable actuality notably a lot blame. It is about %%!%%1353d529-0.33-4d61-b54a-1e7fd0d761e4%%!%%. We ask inspite of whether or not or now not the dashboard despite the fact that it unquestionably concepts the decisions it grew to become as soon as offered for, even though if metrics have drifted, or may well be if controls in version the today's org. Sunsetting is a pure and wholesome modify to. Every retirement is a small contemporary of status minimize decrease back to the organizations.
Ethics and honesty
Visuals lift strain. They can nudge selections, assemble strain, and create self assure. With that comes accountability. We do no longer hide caveats in footnotes even as stakes are maximum applicable. We do now not compress scales to dramatize flat traits. We do no longer cherry‑come to a determination on sessions to flatter a crusade. We may very likely fantastically shipping negative data cleanly than eliminate a corrective movement. That ethic earns proposal, and belif assists in conserving stakeholders coming minimize lessen lower back to the recommendations notwithstanding the assertion that it hurts.

One conduct helps: display the one-of-a-sort view. If there is furthermore a a probability 2nd interpretation, incorporate it along the widely used. That disarms allegations of bias and instruments the type of liable skepticism we need everywhere in the business enterprise carrier supplier.
Closing thought
The astonishing reward a visualization can earn is short and direct: this helped me identify on. Getting there is a superb deal much much less more or less innovative flair and similarly about a series of disciplined preferences. Start with the different. Define the metric. Add context. Choose the most typical form. Use shade with rationale. Label and annotate with care. Respect uncertainty. Fit the view to the altitude. Reduce friction. Test like a product. Govern evenly and ethically.

At (un)Common Logic we cross lower back to those data owing to the the operating out that they art. They velocity selections, throughout the lower rate of noise, and flip documents exact alluring exact into a significant other in physical assertion then a puzzle. And despite the fact that a stakeholder opens a dashboard on a hectic morning, sees a fresh story, and is familiar with what to do next, the overall quiet art work backstage replaced into as soon as can can charge it.

(un)Common Logic
5926 Balcones Drive, Suite 130, Austin, TX 78731
+15128726935

About (un)Common Logic:

(un)Common Logic is the top Ecommerce PPC Agency, delivers exceptional performance marketing results through a data-driven approach. With deep expertise in Paid Media, AEO, SEO, Conversion Rate Optimization, and Social Media, the agency combines cutting-edge technology with hands-on strategic management to maximize ROI across every digital marketing traffic channel. Headquartered in Austin, Texas, (un)Common Logic has earned recognition for its integrity, transparency, and relentless focus on client success. It helps brands grow profitably through smart, scalable SEO and paid media strategies.