By K. Abraham Vaynberg, Tom Bidwell, Griffin Gappert, Kent Maghacut, Joseph Ambrosi, and Matt Lano
Ashland Specialty Ingredients, Ashland Inc.

While paint color can transform the way a room looks, how a paint “feels” is critical to those applying the paint—whether they are professional painters or do-it-yourselfers (DIY). Market research has confirmed this time and again. Manufacturers recognize that the ease or effort of paint application is responsible for the first, and possibly longest lasting, consumer impression and will strongly influence the consumer’s perception of product quality. Test methods for evaluating application feel have been limited in accuracy and reproducibility. Moreover, they rely heavily on the subjective judgment of the evaluator.

Recently, a device called Application Reader Technology™ (ART) has been introduced. The ART device generates and records real-time spatial and force data during the paint roll-out process, objectively measuring application feel. The ART device helps fulfill the need for reliable, reproducible, quantitative data that subjective evaluation techniques have left unmet.

Driving the patent-pending ART device is the understanding that the changes in paint formulation cause subtle changes in the paint rollout process. By recording special and force data generated during application, the ART device objectively records the changes from paint formulation differences and provides direct information and insight into consumer preferences.

In a study using the ART device, 11 participants tested 16 commercially available architectural paints. The research showed that, on average, the panel preferred paints manufactured for the DIY market over those sold as contractor paints.

Equally important, the study found that neither paint rheology nor applied properties produce significant correlations with consumer preferences and, therefore, offer no basis for an a priori prediction of consumer preference. The ART device, on the other hand, produced a number of statistically significant and strong correlations between ART device-measurable parameters and consumer preferences. The strong correlations validate the tool’s ability to predict consumer preference, and warrant further advancement and development of the ART device technology.

Evaluation Background

Paints (all eggshell) from the leading four paint manufacturers were used in the study. Each manufacturer was represented with four paints, two intended for the DIY market and two intended for the contractor market. Each pair consisted of one “premium” paint and one “economy” paint, as designated by the manufacturer. All 16 paints were purchased in retail stores in 5-gal pails.

Since applied hiding is an important consideration affecting the painting process, the DIY paints were tinted light blue. DIY paints were applied on top of contractor paints or vice versa.


Table 1 lists the paints by manufacturer, intended market, price point ($/gal), and name, by paint ID. The last column of Table 1 contains the code for each paint according to the following convention: the first letter represents the manufacturer, the second letter represents DIY or contractor, and the numbers represent the retail price per gallon.


Paint Application

The participants painted on a commercially available section of drywall divided into four 7.5 ft x 10 ft sections. Each section was painted with a single paint listed in Table 1.


To provide color contrast between the paints, researchers tinted the DIY paints at 0.5 oz/gal with Phthalo Blue. Painting participants alternated randomly selected sets of four DIY or contractor paints.

The painters used 9 in. rollers with 3/8 in. nap.  Research assistants did the initial roller breaking and loading to ensure that the painters, with their varying levels of skill, did not impact the process.

Paint loading and spread rate were measured by a scale. Upon completion of the experiment, each participant filled out an evaluation form, described in the next section.

The Evaluation

Table 2 contains the paint evaluation form, which was used to survey participants at the end of each paint application. Participants responded to statements about various aspects of the paint and the application by expressing the extent of their agreement, where “Strongly Disagree” equaled 1 and “Strongly Agree” equaled 9.


The ART device digitally records the painting process, including the position of the roller as a function of time, along with the forces exerted during the painting process. The collected data was used to calculate parameters summarized in Table 3.

To consistently describe dependencies among the various data in this report, Table 4 defines the correlations. Both “Strong” and “Significant” correlations refer to statistically significant correlations.

Discussion of Results

Table 5 shows the results of the painting experience evaluation, which address various attributes of the painting process.

Steady Release describes the ease of paint transfer from roller to the substrate.

Ease of Transfer describes the painter’s physical effort to transfer paint to the substrate.

Roller Pickup describes the amount of paint a roller cover can carry, thus providing for an efficient painting experience.

Wet Film rating covers the combination of wet film hiding and smoothness,

Minimum Reworking refers to the need to go over the same area to deliver better surface finish.

The last two attributes address the user’s perception of paint quality and if he or she would choose the paint.

Participants assigned each attribute a value from 1 to 9. Table 5 contains the painting attribute values averaged over 11 volunteers. The table is color-coded; the lowest values are highlighted red and the highest values green.

Figure 1 shows a correlation between paint price and personal preference. The data follows a clear trend that closely correlates price and the average personal preference score. The data shows three outliers: Ad25, Cd21, and Dd69. The high personal preference scores and low prices indicate that both Ad25 and Cd21 are underpriced. By contrast, Dd69 could be considered overpriced due to its high price and relatively low personal preference score.

This study also sought to define which paint perception attributes most impact overall preference. Table 6 shows the correlations between personal preference and all other perception evaluation parameters. The parameters are arranged by decreasing correlation significance, from most to least. A value above 0.8 is considered statistically very significant. As expected, quality and personal preference are the most highly correlated. This is followed by reworking and wet film appearance. Roller pickup, on the other hand, although demonstrating a strong correlation with personal preference, showed the lowest correlation among the parameters.

Paint Characterization: Applied Properties and Rheology

All 16 paints have been characterized using standard paint characterization procedures. The results of paint characterization are summarized in Table 7. These included KU, ICI, and Brookfield viscosities, leveling, sag, and open time. As expected, paint KU viscosities are near 100KU, ranging between 86KU (Bc42) and 111.2KU (Bd42). ICI viscosities range between 0.60P (Ac17) and 1.43P (Cd21).

Leveling values range between 3 and 9. Bd42 and Dc40 measured leveling of 3. A significant number of paints measured leveling of 9. (In fact, 50% of all paints had leveling values of 9.)  Sag values varied between 6 (Ad42) and 40 for Dc41. Finally, open time ranged between 0 and 8 min. The longest open time, 8 min, was observed with Ad25 paint.

Table 7 offers a unique opportunity to observe similarities and differences between paints intended for the contractor and DIY markets. There is no difference in KU viscosities. Several properties tend to be higher in DIY formulations compared to the contractor formulations, including ICI viscosity, leveling, and open time.

Rheological characterization of the paints was carried out using 40 mm parallel plates. Steady shear viscosity was measured between the shear rates of 0.01 and 1000 s-1. Slope values of viscosity vs shear rate on log scale were calculated between 0.1 and 100 s-1. Dynamic properties of the paints have been measured in linear viscoelastic region between frequencies 0.1 and 100 rad/s. For statistical analysis, the data set was trimmed to include only values corresponding to the rates at each order of magnitude (i.e., at 0.01, 0.1 s-1, etc.). The compilation of rheology data can be found in Tables 8 and 9.

Correlations with Personal Preference

Understanding the relationships between personal preferences and the applied and rheological characteristics of paints can help develop predictive tools for consumer perception. This may reduce or eliminate the industry’s reliance on subjective panel testing.

Personal Preference Correlations with Applied Paint Characteristics

Applied paint properties, such as leveling or ICI viscosity, are commonly used as targets in formulating paints. It is, therefore, important to explore how these parameters correlate with consumer preference.

Table 10 shows pairwise correlations between the average personal preference and the various applied paint properties. The data show that open time is the only parameter yielding significant positive correlation with personal preference.

Open time is the measure of how long the paint remains sufficiently fluid to allow the painter to correct or smooth out imperfections during the painting process. Since the painters in our panel steadily painted each 10-ft section from edge to edge without engaging in any touch-ups or paint smoothing, it is difficult to understand the importance and relevance of the open time to personal preference.

One possible explanation could be that all DIY formulations have higher open time values. Since DIY paints are generally preferred by our panel over the contractor formulations (Table 4), the observed correlation with the open time could be circumstantial.

Table 10 also shows several weak correlations between personal preference and applied paint properties. There is a weak positive correlation with ICI viscosities, suggesting that higher ICI measurements positively affect consumer preference. There is also a weak negative correlation with Brookfield low shear viscosity.

Personal Preferences vs Paint Rheology

Next, we consider the correlation between personal preference and rheological paint data, as summarized in Table 11. The data yield no strong statistical correlations. Note that the rheological characterization consists of both dynamic and steady shear data. Furthermore, rheological data also contains the slope in log-log representation of viscosity vs shear rate and is the measure of paint pseudoplasticity.

Table 11 shows only weak correlations between personal preference and rheological characteristics of the paints. There is a positive weak correlation with tanδ at 100 rad/s, suggesting the preference for paints with higher elastic characteristics at higher frequency.  There is also a weak negative correlation with the shear viscosity at 1 s-1, an intermediate shear rate. Understanding the nature of these relationships is difficult. Surprisingly, there is no correlation between personal preference and a paint’s shear thinning characteristics, where the correlation parameter is only –0.38.

Personal Preference vs ART Device Data

Table 12 summarizes the correlations between the mean personal preference and various measured ART device parameters arranged in the order of the decreasing value of the correlation parameter. Unlike applied paint characteristics and rheology data, ART device data show stronger statistical correlations with the personal preference.

Table 12 shows strong negative correlations between the personal preference and the total number of dips (–0.88, not shown in the table) and length rolled. The negative correlation indicates that the shorter the distance, the fewer dips are required to cover the substrate, and the more likely the paint is going to be liked by a consumer.

The number of strokes, painting time, effort, and average velocity all show significant negative correlations with personal preference. The correlation indicates consumers like fewer strokes, shorter painting time, and less effort. Velocity also yields negative significant correlation, suggesting that when a consumer likes a paint, he or she may paint slower.

Summary and Conclusion

This article summarizes a large body of data based on 16 commercial paint formulations. It encompasses data on consumer preference, ART device parameters, rheology, and applied paint characteristics.

Consumer preference studies are expensive, time-consuming, and ultimately produce results with large variability. The purpose of this study was to establish how personal preferences can be predicted with a degree of certainty without a full consumer panel study. It also sought to identify paint characteristics formulators can embrace to produce paints that are likely to strongly appeal to consumer preferences.

The findings show that both rheology and applied paint characteristics produce only weak correlations with personal preference, and, therefore, cannot be used as independent indicators of consumer preference. The few weak correlations observed were the negative correlations with Brookfield viscosity and steady shear viscosity at 1 s-1. Positive weak correlations were observed with ICI viscosity and tanδ at 100 rad/s. These correlations are not intuitive, and their significance is not explored further.

The Application Reader Technology device, on the other hand, unequivocally demonstrated its value by generating parameters that yielded strong or significant correlations with personal preference. ART data output can be used to predict paints likely to be preferred by consumers. As demonstrated, the paints that take the least number of dips per given substrate size, that require the least distance to roll, the fewest strokes, the shortest time to paint with the slowest velocity, and the least effort are expected to be preferred by a consumer panel.